id
stringlengths 9
16
| submitter
stringlengths 3
64
⌀ | authors
stringlengths 5
6.63k
| title
stringlengths 7
245
| comments
stringlengths 1
482
⌀ | journal-ref
stringlengths 4
382
⌀ | doi
stringlengths 9
151
⌀ | report-no
stringclasses 984
values | categories
stringlengths 5
108
| license
stringclasses 9
values | abstract
stringlengths 83
3.41k
| versions
listlengths 1
20
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-11 00:00:00
| authors_parsed
sequencelengths 1
427
| prompt
stringlengths 166
3.49k
| label
stringclasses 2
values | prob
float64 0.5
0.98
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1611.00889 | Kasra Khosoussi | Kasra Khosoussi, Gaurav S. Sukhatme, Shoudong Huang, Gamini
Dissanayake | Designing Sparse Reliable Pose-Graph SLAM: A Graph-Theoretic Approach | WAFR 2016 | null | null | null | cs.RO cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we aim to design sparse D-optimal (determinantoptimal)
pose-graph SLAM problems through the synthesis of sparse graphs with the
maximum weighted number of spanning trees. Characterizing graphs with the
maximum number of spanning trees is an open problem in general. To tackle this
problem, several new theoretical results are established in this paper,
including the monotone log-submodularity of the weighted number of spanning
trees. By exploiting these structures, we design a complementary pair of
near-optimal efficient approximation algorithms with provable guarantees. Our
theoretical results are validated using random graphs and a publicly available
pose-graph SLAM dataset.
| [
{
"version": "v1",
"created": "Thu, 3 Nov 2016 05:52:37 GMT"
}
] | 2016-11-04T00:00:00 | [
[
"Khosoussi",
"Kasra",
""
],
[
"Sukhatme",
"Gaurav S.",
""
],
[
"Huang",
"Shoudong",
""
],
[
"Dissanayake",
"Gamini",
""
]
] | TITLE: Designing Sparse Reliable Pose-Graph SLAM: A Graph-Theoretic Approach
ABSTRACT: In this paper, we aim to design sparse D-optimal (determinantoptimal)
pose-graph SLAM problems through the synthesis of sparse graphs with the
maximum weighted number of spanning trees. Characterizing graphs with the
maximum number of spanning trees is an open problem in general. To tackle this
problem, several new theoretical results are established in this paper,
including the monotone log-submodularity of the weighted number of spanning
trees. By exploiting these structures, we design a complementary pair of
near-optimal efficient approximation algorithms with provable guarantees. Our
theoretical results are validated using random graphs and a publicly available
pose-graph SLAM dataset.
| new_dataset | 0.932883 |
1506.05532 | Salman Khan Mr. | Munawar Hayat, Salman H. Khan, Mohammed Bennamoun, Senjian An | A Spatial Layout and Scale Invariant Feature Representation for Indoor
Scene Classification | null | null | 10.1109/TIP.2016.2599292 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unlike standard object classification, where the image to be classified
contains one or multiple instances of the same object, indoor scene
classification is quite different since the image consists of multiple distinct
objects. Further, these objects can be of varying sizes and are present across
numerous spatial locations in different layouts. For automatic indoor scene
categorization, large scale spatial layout deformations and scale variations
are therefore two major challenges and the design of rich feature descriptors
which are robust to these challenges is still an open problem. This paper
introduces a new learnable feature descriptor called "spatial layout and scale
invariant convolutional activations" to deal with these challenges. For this
purpose, a new Convolutional Neural Network architecture is designed which
incorporates a novel 'Spatially Unstructured' layer to introduce robustness
against spatial layout deformations. To achieve scale invariance, we present a
pyramidal image representation. For feasible training of the proposed network
for images of indoor scenes, the paper proposes a new methodology which
efficiently adapts a trained network model (on a large scale data) for our task
with only a limited amount of available training data. Compared with existing
state of the art, the proposed approach achieves a relative performance
improvement of 3.2%, 3.8%, 7.0%, 11.9% and 2.1% on MIT-67, Scene-15, Sports-8,
Graz-02 and NYU datasets respectively.
| [
{
"version": "v1",
"created": "Thu, 18 Jun 2015 02:11:37 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Aug 2015 04:01:11 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Hayat",
"Munawar",
""
],
[
"Khan",
"Salman H.",
""
],
[
"Bennamoun",
"Mohammed",
""
],
[
"An",
"Senjian",
""
]
] | TITLE: A Spatial Layout and Scale Invariant Feature Representation for Indoor
Scene Classification
ABSTRACT: Unlike standard object classification, where the image to be classified
contains one or multiple instances of the same object, indoor scene
classification is quite different since the image consists of multiple distinct
objects. Further, these objects can be of varying sizes and are present across
numerous spatial locations in different layouts. For automatic indoor scene
categorization, large scale spatial layout deformations and scale variations
are therefore two major challenges and the design of rich feature descriptors
which are robust to these challenges is still an open problem. This paper
introduces a new learnable feature descriptor called "spatial layout and scale
invariant convolutional activations" to deal with these challenges. For this
purpose, a new Convolutional Neural Network architecture is designed which
incorporates a novel 'Spatially Unstructured' layer to introduce robustness
against spatial layout deformations. To achieve scale invariance, we present a
pyramidal image representation. For feasible training of the proposed network
for images of indoor scenes, the paper proposes a new methodology which
efficiently adapts a trained network model (on a large scale data) for our task
with only a limited amount of available training data. Compared with existing
state of the art, the proposed approach achieves a relative performance
improvement of 3.2%, 3.8%, 7.0%, 11.9% and 2.1% on MIT-67, Scene-15, Sports-8,
Graz-02 and NYU datasets respectively.
| no_new_dataset | 0.950595 |
1511.08058 | Yanwei Pang | Jiale Cao, Yanwei Pang, and Xuelong Li | Pedestrian Detection Inspired by Appearance Constancy and Shape Symmetry | 9 pages,17 figures | null | 10.1109/TIP.2016.2609807 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The discrimination and simplicity of features are very important for
effective and efficient pedestrian detection. However, most state-of-the-art
methods are unable to achieve good tradeoff between accuracy and efficiency.
Inspired by some simple inherent attributes of pedestrians (i.e., appearance
constancy and shape symmetry), we propose two new types of non-neighboring
features (NNF): side-inner difference features (SIDF) and symmetrical
similarity features (SSF). SIDF can characterize the difference between the
background and pedestrian and the difference between the pedestrian contour and
its inner part. SSF can capture the symmetrical similarity of pedestrian shape.
However, it's difficult for neighboring features to have such above
characterization abilities. Finally, we propose to combine both non-neighboring
and neighboring features for pedestrian detection. It's found that
non-neighboring features can further decrease the average miss rate by 4.44%.
Experimental results on INRIA and Caltech pedestrian datasets demonstrate the
effectiveness and efficiency of the proposed method. Compared to the
state-of-the-art methods without using CNN, our method achieves the best
detection performance on Caltech, outperforming the second best method (i.e.,
Checkboards) by 1.63%.
| [
{
"version": "v1",
"created": "Wed, 25 Nov 2015 13:49:13 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Cao",
"Jiale",
""
],
[
"Pang",
"Yanwei",
""
],
[
"Li",
"Xuelong",
""
]
] | TITLE: Pedestrian Detection Inspired by Appearance Constancy and Shape Symmetry
ABSTRACT: The discrimination and simplicity of features are very important for
effective and efficient pedestrian detection. However, most state-of-the-art
methods are unable to achieve good tradeoff between accuracy and efficiency.
Inspired by some simple inherent attributes of pedestrians (i.e., appearance
constancy and shape symmetry), we propose two new types of non-neighboring
features (NNF): side-inner difference features (SIDF) and symmetrical
similarity features (SSF). SIDF can characterize the difference between the
background and pedestrian and the difference between the pedestrian contour and
its inner part. SSF can capture the symmetrical similarity of pedestrian shape.
However, it's difficult for neighboring features to have such above
characterization abilities. Finally, we propose to combine both non-neighboring
and neighboring features for pedestrian detection. It's found that
non-neighboring features can further decrease the average miss rate by 4.44%.
Experimental results on INRIA and Caltech pedestrian datasets demonstrate the
effectiveness and efficiency of the proposed method. Compared to the
state-of-the-art methods without using CNN, our method achieves the best
detection performance on Caltech, outperforming the second best method (i.e.,
Checkboards) by 1.63%.
| no_new_dataset | 0.946843 |
1606.03757 | Brendon Brewer | Brendon J. Brewer and Daniel Foreman-Mackey | DNest4: Diffusive Nested Sampling in C++ and Python | Submitted. 33 pages, 9 figures. v2 removed a duplicated figure, v3
added a comparison to other packages, v4 fixed a few minor issues | null | null | null | stat.CO astro-ph.IM physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In probabilistic (Bayesian) inferences, we typically want to compute
properties of the posterior distribution, describing knowledge of unknown
quantities in the context of a particular dataset and the assumed prior
information. The marginal likelihood, also known as the "evidence", is a key
quantity in Bayesian model selection. The Diffusive Nested Sampling algorithm,
a variant of Nested Sampling, is a powerful tool for generating posterior
samples and estimating marginal likelihoods. It is effective at solving complex
problems including many where the posterior distribution is multimodal or has
strong dependencies between variables. DNest4 is an open source (MIT licensed),
multi-threaded implementation of this algorithm in C++11, along with associated
utilities including: i) RJObject, a class template for finite mixture models,
(ii) A Python package allowing basic use without C++ coding, and iii)
Experimental support for models implemented in Julia. In this paper we
demonstrate DNest4 usage through examples including simple Bayesian data
analysis, finite mixture models, and Approximate Bayesian Computation.
| [
{
"version": "v1",
"created": "Sun, 12 Jun 2016 19:21:30 GMT"
},
{
"version": "v2",
"created": "Tue, 14 Jun 2016 00:57:29 GMT"
},
{
"version": "v3",
"created": "Tue, 26 Jul 2016 02:22:54 GMT"
},
{
"version": "v4",
"created": "Wed, 2 Nov 2016 01:15:31 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Brewer",
"Brendon J.",
""
],
[
"Foreman-Mackey",
"Daniel",
""
]
] | TITLE: DNest4: Diffusive Nested Sampling in C++ and Python
ABSTRACT: In probabilistic (Bayesian) inferences, we typically want to compute
properties of the posterior distribution, describing knowledge of unknown
quantities in the context of a particular dataset and the assumed prior
information. The marginal likelihood, also known as the "evidence", is a key
quantity in Bayesian model selection. The Diffusive Nested Sampling algorithm,
a variant of Nested Sampling, is a powerful tool for generating posterior
samples and estimating marginal likelihoods. It is effective at solving complex
problems including many where the posterior distribution is multimodal or has
strong dependencies between variables. DNest4 is an open source (MIT licensed),
multi-threaded implementation of this algorithm in C++11, along with associated
utilities including: i) RJObject, a class template for finite mixture models,
(ii) A Python package allowing basic use without C++ coding, and iii)
Experimental support for models implemented in Julia. In this paper we
demonstrate DNest4 usage through examples including simple Bayesian data
analysis, finite mixture models, and Approximate Bayesian Computation.
| no_new_dataset | 0.949902 |
1607.08811 | Pedro Herruzo | Pedro Herruzo, Marc Bola\~nos and Petia Radeva | Can a CNN Recognize Catalan Diet? | 9 pages, 6 figures, 6 tables | null | 10.1063/1.4964956 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nowadays, we can find several diseases related to the unhealthy diet habits
of the population, such as diabetes, obesity, anemia, bulimia and anorexia. In
many cases, these diseases are related to the food consumption of people.
Mediterranean diet is scientifically known as a healthy diet that helps to
prevent many metabolic diseases. In particular, our work focuses on the
recognition of Mediterranean food and dishes. The development of this
methodology would allow to analise the daily habits of users with wearable
cameras, within the topic of lifelogging. By using automatic mechanisms we
could build an objective tool for the analysis of the patient's behaviour,
allowing specialists to discover unhealthy food patterns and understand the
user's lifestyle.
With the aim to automatically recognize a complete diet, we introduce a
challenging multi-labeled dataset related to Mediterranean diet called FoodCAT.
The first type of label provided consists of 115 food classes with an average
of 400 images per dish, and the second one consists of 12 food categories with
an average of 3800 pictures per class. This dataset will serve as a basis for
the development of automatic diet recognition. In this context, deep learning
and more specifically, Convolutional Neural Networks (CNNs), currently are
state-of-the-art methods for automatic food recognition. In our work, we
compare several architectures for image classification, with the purpose of
diet recognition. Applying the best model for recognising food categories, we
achieve a top-1 accuracy of 72.29\%, and top-5 of 97.07\%. In a complete diet
recognition of dishes from Mediterranean diet, enlarged with the Food-101
dataset for international dishes recognition, we achieve a top-1 accuracy of
68.07\%, and top-5 of 89.53\%, for a total of 115+101 food classes.
| [
{
"version": "v1",
"created": "Fri, 29 Jul 2016 13:55:21 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Herruzo",
"Pedro",
""
],
[
"Bolaños",
"Marc",
""
],
[
"Radeva",
"Petia",
""
]
] | TITLE: Can a CNN Recognize Catalan Diet?
ABSTRACT: Nowadays, we can find several diseases related to the unhealthy diet habits
of the population, such as diabetes, obesity, anemia, bulimia and anorexia. In
many cases, these diseases are related to the food consumption of people.
Mediterranean diet is scientifically known as a healthy diet that helps to
prevent many metabolic diseases. In particular, our work focuses on the
recognition of Mediterranean food and dishes. The development of this
methodology would allow to analise the daily habits of users with wearable
cameras, within the topic of lifelogging. By using automatic mechanisms we
could build an objective tool for the analysis of the patient's behaviour,
allowing specialists to discover unhealthy food patterns and understand the
user's lifestyle.
With the aim to automatically recognize a complete diet, we introduce a
challenging multi-labeled dataset related to Mediterranean diet called FoodCAT.
The first type of label provided consists of 115 food classes with an average
of 400 images per dish, and the second one consists of 12 food categories with
an average of 3800 pictures per class. This dataset will serve as a basis for
the development of automatic diet recognition. In this context, deep learning
and more specifically, Convolutional Neural Networks (CNNs), currently are
state-of-the-art methods for automatic food recognition. In our work, we
compare several architectures for image classification, with the purpose of
diet recognition. Applying the best model for recognising food categories, we
achieve a top-1 accuracy of 72.29\%, and top-5 of 97.07\%. In a complete diet
recognition of dishes from Mediterranean diet, enlarged with the Food-101
dataset for international dishes recognition, we achieve a top-1 accuracy of
68.07\%, and top-5 of 89.53\%, for a total of 115+101 food classes.
| new_dataset | 0.967287 |
1609.02077 | Guanbin Li | Guanbin Li and Yizhou Yu | Visual Saliency Detection Based on Multiscale Deep CNN Features | Accepted for publication in IEEE Transactions on Image Processing | null | 10.1109/TIP.2016.2602079 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual saliency is a fundamental problem in both cognitive and computational
sciences, including computer vision. In this paper, we discover that a
high-quality visual saliency model can be learned from multiscale features
extracted using deep convolutional neural networks (CNNs), which have had many
successes in visual recognition tasks. For learning such saliency models, we
introduce a neural network architecture, which has fully connected layers on
top of CNNs responsible for feature extraction at three different scales. The
penultimate layer of our neural network has been confirmed to be a
discriminative high-level feature vector for saliency detection, which we call
deep contrast feature. To generate a more robust feature, we integrate
handcrafted low-level features with our deep contrast feature. To promote
further research and evaluation of visual saliency models, we also construct a
new large database of 4447 challenging images and their pixelwise saliency
annotations. Experimental results demonstrate that our proposed method is
capable of achieving state-of-the-art performance on all public benchmarks,
improving the F- measure by 6.12% and 10.0% respectively on the DUT-OMRON
dataset and our new dataset (HKU-IS), and lowering the mean absolute error by
9% and 35.3% respectively on these two datasets.
| [
{
"version": "v1",
"created": "Wed, 7 Sep 2016 17:13:16 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Li",
"Guanbin",
""
],
[
"Yu",
"Yizhou",
""
]
] | TITLE: Visual Saliency Detection Based on Multiscale Deep CNN Features
ABSTRACT: Visual saliency is a fundamental problem in both cognitive and computational
sciences, including computer vision. In this paper, we discover that a
high-quality visual saliency model can be learned from multiscale features
extracted using deep convolutional neural networks (CNNs), which have had many
successes in visual recognition tasks. For learning such saliency models, we
introduce a neural network architecture, which has fully connected layers on
top of CNNs responsible for feature extraction at three different scales. The
penultimate layer of our neural network has been confirmed to be a
discriminative high-level feature vector for saliency detection, which we call
deep contrast feature. To generate a more robust feature, we integrate
handcrafted low-level features with our deep contrast feature. To promote
further research and evaluation of visual saliency models, we also construct a
new large database of 4447 challenging images and their pixelwise saliency
annotations. Experimental results demonstrate that our proposed method is
capable of achieving state-of-the-art performance on all public benchmarks,
improving the F- measure by 6.12% and 10.0% respectively on the DUT-OMRON
dataset and our new dataset (HKU-IS), and lowering the mean absolute error by
9% and 35.3% respectively on these two datasets.
| new_dataset | 0.955569 |
1610.07184 | Soumitra Pal | Soumitra Pal, Tingyang Xu, Tianbao Yang, Sanguthevar Rajasekaran,
Jinbo Bi | Hybrid-DCA: A Double Asynchronous Approach for Stochastic Dual
Coordinate Ascent | null | null | null | null | cs.DC math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In prior works, stochastic dual coordinate ascent (SDCA) has been
parallelized in a multi-core environment where the cores communicate through
shared memory, or in a multi-processor distributed memory environment where the
processors communicate through message passing. In this paper, we propose a
hybrid SDCA framework for multi-core clusters, the most common high performance
computing environment that consists of multiple nodes each having multiple
cores and its own shared memory. We distribute data across nodes where each
node solves a local problem in an asynchronous parallel fashion on its cores,
and then the local updates are aggregated via an asynchronous across-node
update scheme. The proposed double asynchronous method converges to a global
solution for $L$-Lipschitz continuous loss functions, and at a linear
convergence rate if a smooth convex loss function is used. Extensive empirical
comparison has shown that our algorithm scales better than the best known
shared-memory methods and runs faster than previous distributed-memory methods.
Big datasets, such as one of 280 GB from the LIBSVM repository, cannot be
accommodated on a single node and hence cannot be solved by a parallel
algorithm. For such a dataset, our hybrid algorithm takes 30 seconds to achieve
a duality gap of $10^{-6}$ on 16 nodes each using 8 cores, which is
significantly faster than the best known distributed algorithms, such as
CoCoA+, that take more than 300 seconds on 16 nodes.
| [
{
"version": "v1",
"created": "Sun, 23 Oct 2016 15:17:43 GMT"
},
{
"version": "v2",
"created": "Wed, 2 Nov 2016 17:50:50 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Pal",
"Soumitra",
""
],
[
"Xu",
"Tingyang",
""
],
[
"Yang",
"Tianbao",
""
],
[
"Rajasekaran",
"Sanguthevar",
""
],
[
"Bi",
"Jinbo",
""
]
] | TITLE: Hybrid-DCA: A Double Asynchronous Approach for Stochastic Dual
Coordinate Ascent
ABSTRACT: In prior works, stochastic dual coordinate ascent (SDCA) has been
parallelized in a multi-core environment where the cores communicate through
shared memory, or in a multi-processor distributed memory environment where the
processors communicate through message passing. In this paper, we propose a
hybrid SDCA framework for multi-core clusters, the most common high performance
computing environment that consists of multiple nodes each having multiple
cores and its own shared memory. We distribute data across nodes where each
node solves a local problem in an asynchronous parallel fashion on its cores,
and then the local updates are aggregated via an asynchronous across-node
update scheme. The proposed double asynchronous method converges to a global
solution for $L$-Lipschitz continuous loss functions, and at a linear
convergence rate if a smooth convex loss function is used. Extensive empirical
comparison has shown that our algorithm scales better than the best known
shared-memory methods and runs faster than previous distributed-memory methods.
Big datasets, such as one of 280 GB from the LIBSVM repository, cannot be
accommodated on a single node and hence cannot be solved by a parallel
algorithm. For such a dataset, our hybrid algorithm takes 30 seconds to achieve
a duality gap of $10^{-6}$ on 16 nodes each using 8 cores, which is
significantly faster than the best known distributed algorithms, such as
CoCoA+, that take more than 300 seconds on 16 nodes.
| no_new_dataset | 0.944022 |
1610.09650 | Bharat Sau | Bharat Bhusan Sau and Vineeth N. Balasubramanian | Deep Model Compression: Distilling Knowledge from Noisy Teachers | 9 pages, 3 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The remarkable successes of deep learning models across various applications
have resulted in the design of deeper networks that can solve complex problems.
However, the increasing depth of such models also results in a higher storage
and runtime complexity, which restricts the deployability of such very deep
models on mobile and portable devices, which have limited storage and battery
capacity. While many methods have been proposed for deep model compression in
recent years, almost all of them have focused on reducing storage complexity.
In this work, we extend the teacher-student framework for deep model
compression, since it has the potential to address runtime and train time
complexity too. We propose a simple methodology to include a noise-based
regularizer while training the student from the teacher, which provides a
healthy improvement in the performance of the student network. Our experiments
on the CIFAR-10, SVHN and MNIST datasets show promising improvement, with the
best performance on the CIFAR-10 dataset. We also conduct a comprehensive
empirical evaluation of the proposed method under related settings on the
CIFAR-10 dataset to show the promise of the proposed approach.
| [
{
"version": "v1",
"created": "Sun, 30 Oct 2016 13:54:39 GMT"
},
{
"version": "v2",
"created": "Wed, 2 Nov 2016 16:32:23 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Sau",
"Bharat Bhusan",
""
],
[
"Balasubramanian",
"Vineeth N.",
""
]
] | TITLE: Deep Model Compression: Distilling Knowledge from Noisy Teachers
ABSTRACT: The remarkable successes of deep learning models across various applications
have resulted in the design of deeper networks that can solve complex problems.
However, the increasing depth of such models also results in a higher storage
and runtime complexity, which restricts the deployability of such very deep
models on mobile and portable devices, which have limited storage and battery
capacity. While many methods have been proposed for deep model compression in
recent years, almost all of them have focused on reducing storage complexity.
In this work, we extend the teacher-student framework for deep model
compression, since it has the potential to address runtime and train time
complexity too. We propose a simple methodology to include a noise-based
regularizer while training the student from the teacher, which provides a
healthy improvement in the performance of the student network. Our experiments
on the CIFAR-10, SVHN and MNIST datasets show promising improvement, with the
best performance on the CIFAR-10 dataset. We also conduct a comprehensive
empirical evaluation of the proposed method under related settings on the
CIFAR-10 dataset to show the promise of the proposed approach.
| no_new_dataset | 0.944382 |
1610.09996 | Yang Yu | Yang Yu, Wei Zhang, Kazi Hasan, Mo Yu, Bing Xiang, Bowen Zhou | End-to-End Answer Chunk Extraction and Ranking for Reading Comprehension | Submitted to AAAI | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes dynamic chunk reader (DCR), an end-to-end neural reading
comprehension (RC) model that is able to extract and rank a set of answer
candidates from a given document to answer questions. DCR is able to predict
answers of variable lengths, whereas previous neural RC models primarily
focused on predicting single tokens or entities. DCR encodes a document and an
input question with recurrent neural networks, and then applies a word-by-word
attention mechanism to acquire question-aware representations for the document,
followed by the generation of chunk representations and a ranking module to
propose the top-ranked chunk as the answer. Experimental results show that DCR
achieves state-of-the-art exact match and F1 scores on the SQuAD dataset.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 16:14:08 GMT"
},
{
"version": "v2",
"created": "Wed, 2 Nov 2016 17:55:32 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Yu",
"Yang",
""
],
[
"Zhang",
"Wei",
""
],
[
"Hasan",
"Kazi",
""
],
[
"Yu",
"Mo",
""
],
[
"Xiang",
"Bing",
""
],
[
"Zhou",
"Bowen",
""
]
] | TITLE: End-to-End Answer Chunk Extraction and Ranking for Reading Comprehension
ABSTRACT: This paper proposes dynamic chunk reader (DCR), an end-to-end neural reading
comprehension (RC) model that is able to extract and rank a set of answer
candidates from a given document to answer questions. DCR is able to predict
answers of variable lengths, whereas previous neural RC models primarily
focused on predicting single tokens or entities. DCR encodes a document and an
input question with recurrent neural networks, and then applies a word-by-word
attention mechanism to acquire question-aware representations for the document,
followed by the generation of chunk representations and a ranking module to
propose the top-ranked chunk as the answer. Experimental results show that DCR
achieves state-of-the-art exact match and F1 scores on the SQuAD dataset.
| no_new_dataset | 0.950595 |
1611.00336 | Andrew Wilson | Andrew Gordon Wilson, Zhiting Hu, Ruslan Salakhutdinov, Eric P. Xing | Stochastic Variational Deep Kernel Learning | 13 pages, 6 tables, 3 figures. Appearing in NIPS 2016 | null | null | null | stat.ML cs.LG stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep kernel learning combines the non-parametric flexibility of kernel
methods with the inductive biases of deep learning architectures. We propose a
novel deep kernel learning model and stochastic variational inference procedure
which generalizes deep kernel learning approaches to enable classification,
multi-task learning, additive covariance structures, and stochastic gradient
training. Specifically, we apply additive base kernels to subsets of output
features from deep neural architectures, and jointly learn the parameters of
the base kernels and deep network through a Gaussian process marginal
likelihood objective. Within this framework, we derive an efficient form of
stochastic variational inference which leverages local kernel interpolation,
inducing points, and structure exploiting algebra. We show improved performance
over stand alone deep networks, SVMs, and state of the art scalable Gaussian
processes on several classification benchmarks, including an airline delay
dataset containing 6 million training points, CIFAR, and ImageNet.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 19:04:47 GMT"
},
{
"version": "v2",
"created": "Wed, 2 Nov 2016 18:06:16 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Wilson",
"Andrew Gordon",
""
],
[
"Hu",
"Zhiting",
""
],
[
"Salakhutdinov",
"Ruslan",
""
],
[
"Xing",
"Eric P.",
""
]
] | TITLE: Stochastic Variational Deep Kernel Learning
ABSTRACT: Deep kernel learning combines the non-parametric flexibility of kernel
methods with the inductive biases of deep learning architectures. We propose a
novel deep kernel learning model and stochastic variational inference procedure
which generalizes deep kernel learning approaches to enable classification,
multi-task learning, additive covariance structures, and stochastic gradient
training. Specifically, we apply additive base kernels to subsets of output
features from deep neural architectures, and jointly learn the parameters of
the base kernels and deep network through a Gaussian process marginal
likelihood objective. Within this framework, we derive an efficient form of
stochastic variational inference which leverages local kernel interpolation,
inducing points, and structure exploiting algebra. We show improved performance
over stand alone deep networks, SVMs, and state of the art scalable Gaussian
processes on several classification benchmarks, including an airline delay
dataset containing 6 million training points, CIFAR, and ImageNet.
| no_new_dataset | 0.948489 |
1611.00379 | Baptiste Caramiaux | Rebecca Fiebrink, Baptiste Caramiaux | The Machine Learning Algorithm as Creative Musical Tool | Pre-print to appear in the Oxford Handbook on Algorithmic Music.
Oxford University Press | null | null | null | cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning is the capacity of a computational system to learn
structures from datasets in order to make predictions on newly seen data. Such
an approach offers a significant advantage in music scenarios in which
musicians can teach the system to learn an idiosyncratic style, or can break
the rules to explore the system's capacity in unexpected ways. In this chapter
we draw on music, machine learning, and human-computer interaction to elucidate
an understanding of machine learning algorithms as creative tools for music and
the sonic arts. We motivate a new understanding of learning algorithms as
human-computer interfaces. We show that, like other interfaces, learning
algorithms can be characterised by the ways their affordances intersect with
goals of human users. We also argue that the nature of interaction between
users and algorithms impacts the usability and usefulness of those algorithms
in profound ways. This human-centred view of machine learning motivates our
concluding discussion of what it means to employ machine learning as a creative
tool.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 20:35:46 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Fiebrink",
"Rebecca",
""
],
[
"Caramiaux",
"Baptiste",
""
]
] | TITLE: The Machine Learning Algorithm as Creative Musical Tool
ABSTRACT: Machine learning is the capacity of a computational system to learn
structures from datasets in order to make predictions on newly seen data. Such
an approach offers a significant advantage in music scenarios in which
musicians can teach the system to learn an idiosyncratic style, or can break
the rules to explore the system's capacity in unexpected ways. In this chapter
we draw on music, machine learning, and human-computer interaction to elucidate
an understanding of machine learning algorithms as creative tools for music and
the sonic arts. We motivate a new understanding of learning algorithms as
human-computer interfaces. We show that, like other interfaces, learning
algorithms can be characterised by the ways their affordances intersect with
goals of human users. We also argue that the nature of interaction between
users and algorithms impacts the usability and usefulness of those algorithms
in profound ways. This human-centred view of machine learning motivates our
concluding discussion of what it means to employ machine learning as a creative
tool.
| no_new_dataset | 0.945551 |
1611.00423 | Haoyu Zhang | Haoyu Zhang, Qin Zhang | Computing Skylines on Distributed Data | null | null | null | null | cs.DB cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we study skyline queries in the distributed computational
model, where we have $s$ remote sites and a central coordinator (the query
node); each site holds a piece of data, and the coordinator wants to compute
the skyline of the union of the $s$ datasets. The computation is in terms of
rounds, and the goal is to minimize both the total communication cost and the
round cost.
Viewing data objects as points in the Euclidean space, we consider both the
horizontal data partition case where each site holds a subset of points, and
the vertical data partition case where each site holds one coordinate of all
the points. We give a set of algorithms that have provable theoretical
guarantees, and complement them with information theoretical lower bounds. We
also demonstrate the superiority of our algorithms over existing heuristics by
an extensive set of experiments on both synthetic and real world datasets.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 23:41:03 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Zhang",
"Haoyu",
""
],
[
"Zhang",
"Qin",
""
]
] | TITLE: Computing Skylines on Distributed Data
ABSTRACT: In this paper we study skyline queries in the distributed computational
model, where we have $s$ remote sites and a central coordinator (the query
node); each site holds a piece of data, and the coordinator wants to compute
the skyline of the union of the $s$ datasets. The computation is in terms of
rounds, and the goal is to minimize both the total communication cost and the
round cost.
Viewing data objects as points in the Euclidean space, we consider both the
horizontal data partition case where each site holds a subset of points, and
the vertical data partition case where each site holds one coordinate of all
the points. We give a set of algorithms that have provable theoretical
guarantees, and complement them with information theoretical lower bounds. We
also demonstrate the superiority of our algorithms over existing heuristics by
an extensive set of experiments on both synthetic and real world datasets.
| no_new_dataset | 0.948251 |
1611.00448 | Hao Wang | Hao Wang, Xingjian Shi, Dit-Yan Yeung | Natural-Parameter Networks: A Class of Probabilistic Neural Networks | To appear at NIPS 2016 | null | null | null | cs.LG cs.AI cs.CL cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural networks (NN) have achieved state-of-the-art performance in various
applications. Unfortunately in applications where training data is
insufficient, they are often prone to overfitting. One effective way to
alleviate this problem is to exploit the Bayesian approach by using Bayesian
neural networks (BNN). Another shortcoming of NN is the lack of flexibility to
customize different distributions for the weights and neurons according to the
data, as is often done in probabilistic graphical models. To address these
problems, we propose a class of probabilistic neural networks, dubbed
natural-parameter networks (NPN), as a novel and lightweight Bayesian treatment
of NN. NPN allows the usage of arbitrary exponential-family distributions to
model the weights and neurons. Different from traditional NN and BNN, NPN takes
distributions as input and goes through layers of transformation before
producing distributions to match the target output distributions. As a Bayesian
treatment, efficient backpropagation (BP) is performed to learn the natural
parameters for the distributions over both the weights and neurons. The output
distributions of each layer, as byproducts, may be used as second-order
representations for the associated tasks such as link prediction. Experiments
on real-world datasets show that NPN can achieve state-of-the-art performance.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 02:32:05 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Wang",
"Hao",
""
],
[
"Shi",
"Xingjian",
""
],
[
"Yeung",
"Dit-Yan",
""
]
] | TITLE: Natural-Parameter Networks: A Class of Probabilistic Neural Networks
ABSTRACT: Neural networks (NN) have achieved state-of-the-art performance in various
applications. Unfortunately in applications where training data is
insufficient, they are often prone to overfitting. One effective way to
alleviate this problem is to exploit the Bayesian approach by using Bayesian
neural networks (BNN). Another shortcoming of NN is the lack of flexibility to
customize different distributions for the weights and neurons according to the
data, as is often done in probabilistic graphical models. To address these
problems, we propose a class of probabilistic neural networks, dubbed
natural-parameter networks (NPN), as a novel and lightweight Bayesian treatment
of NN. NPN allows the usage of arbitrary exponential-family distributions to
model the weights and neurons. Different from traditional NN and BNN, NPN takes
distributions as input and goes through layers of transformation before
producing distributions to match the target output distributions. As a Bayesian
treatment, efficient backpropagation (BP) is performed to learn the natural
parameters for the distributions over both the weights and neurons. The output
distributions of each layer, as byproducts, may be used as second-order
representations for the associated tasks such as link prediction. Experiments
on real-world datasets show that NPN can achieve state-of-the-art performance.
| no_new_dataset | 0.948442 |
1611.00454 | Hao Wang | Hao Wang, Xingjian Shi, Dit-Yan Yeung | Collaborative Recurrent Autoencoder: Recommend while Learning to Fill in
the Blanks | To appear at NIPS 2016 | null | null | null | cs.LG cs.AI cs.CL cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hybrid methods that utilize both content and rating information are commonly
used in many recommender systems. However, most of them use either handcrafted
features or the bag-of-words representation as a surrogate for the content
information but they are neither effective nor natural enough. To address this
problem, we develop a collaborative recurrent autoencoder (CRAE) which is a
denoising recurrent autoencoder (DRAE) that models the generation of content
sequences in the collaborative filtering (CF) setting. The model generalizes
recent advances in recurrent deep learning from i.i.d. input to non-i.i.d.
(CF-based) input and provides a new denoising scheme along with a novel
learnable pooling scheme for the recurrent autoencoder. To do this, we first
develop a hierarchical Bayesian model for the DRAE and then generalize it to
the CF setting. The synergy between denoising and CF enables CRAE to make
accurate recommendations while learning to fill in the blanks in sequences.
Experiments on real-world datasets from different domains (CiteULike and
Netflix) show that, by jointly modeling the order-aware generation of sequences
for the content information and performing CF for the ratings, CRAE is able to
significantly outperform the state of the art on both the recommendation task
based on ratings and the sequence generation task based on content information.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 02:49:44 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Wang",
"Hao",
""
],
[
"Shi",
"Xingjian",
""
],
[
"Yeung",
"Dit-Yan",
""
]
] | TITLE: Collaborative Recurrent Autoencoder: Recommend while Learning to Fill in
the Blanks
ABSTRACT: Hybrid methods that utilize both content and rating information are commonly
used in many recommender systems. However, most of them use either handcrafted
features or the bag-of-words representation as a surrogate for the content
information but they are neither effective nor natural enough. To address this
problem, we develop a collaborative recurrent autoencoder (CRAE) which is a
denoising recurrent autoencoder (DRAE) that models the generation of content
sequences in the collaborative filtering (CF) setting. The model generalizes
recent advances in recurrent deep learning from i.i.d. input to non-i.i.d.
(CF-based) input and provides a new denoising scheme along with a novel
learnable pooling scheme for the recurrent autoencoder. To do this, we first
develop a hierarchical Bayesian model for the DRAE and then generalize it to
the CF setting. The synergy between denoising and CF enables CRAE to make
accurate recommendations while learning to fill in the blanks in sequences.
Experiments on real-world datasets from different domains (CiteULike and
Netflix) show that, by jointly modeling the order-aware generation of sequences
for the content information and performing CF for the ratings, CRAE is able to
significantly outperform the state of the art on both the recommendation task
based on ratings and the sequence generation task based on content information.
| no_new_dataset | 0.951097 |
1611.00457 | Bo Wang | Bo Wang, Yingjun Sun, Yuan Wang | Structure vs. Language: Investigating the Multi-factors of Asymmetric
Opinions on Online Social Interrelationship with a Case Study | null | null | null | null | cs.SI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Though current researches often study the properties of online social
relationship from an objective view, we also need to understand individuals'
subjective opinions on their interrelationships in social computing studies.
Inspired by the theories from sociolinguistics, the latest work indicates that
interactive language can reveal individuals' asymmetric opinions on their
interrelationship. In this work, in order to explain the opinions' asymmetry on
interrelationship with more latent factors, we extend the investigation from
single relationship to the structural context in online social network. We
analyze the correlation between interactive language features and the
structural context of interrelationships. The structural context of vertex,
edges and triangles in social network are considered. With statistical analysis
on Enron email dataset, we find that individuals' opinions (measured by
interactive language features) on their interrelationship are related to some
of their important structural context in social network. This result can help
us to understand and measure the individuals' opinions on their
interrelationship with more intrinsic information.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 03:11:10 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Wang",
"Bo",
""
],
[
"Sun",
"Yingjun",
""
],
[
"Wang",
"Yuan",
""
]
] | TITLE: Structure vs. Language: Investigating the Multi-factors of Asymmetric
Opinions on Online Social Interrelationship with a Case Study
ABSTRACT: Though current researches often study the properties of online social
relationship from an objective view, we also need to understand individuals'
subjective opinions on their interrelationships in social computing studies.
Inspired by the theories from sociolinguistics, the latest work indicates that
interactive language can reveal individuals' asymmetric opinions on their
interrelationship. In this work, in order to explain the opinions' asymmetry on
interrelationship with more latent factors, we extend the investigation from
single relationship to the structural context in online social network. We
analyze the correlation between interactive language features and the
structural context of interrelationships. The structural context of vertex,
edges and triangles in social network are considered. With statistical analysis
on Enron email dataset, we find that individuals' opinions (measured by
interactive language features) on their interrelationship are related to some
of their important structural context in social network. This result can help
us to understand and measure the individuals' opinions on their
interrelationship with more intrinsic information.
| no_new_dataset | 0.941331 |
1611.00468 | Xiao Chu | Xiao Chu, Wanli Ouyang, Hongsheng Li and Xiaogang Wang | CRF-CNN: Modeling Structured Information in Human Pose Estimation | NIPS | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep convolutional neural networks (CNN) have achieved great success. On the
other hand, modeling structural information has been proved critical in many
vision problems. It is of great interest to integrate them effectively. In a
classical neural network, there is no message passing between neurons in the
same layer. In this paper, we propose a CRF-CNN framework which can
simultaneously model structural information in both output and hidden feature
layers in a probabilistic way, and it is applied to human pose estimation. A
message passing scheme is proposed, so that in various layers each body joint
receives messages from all the others in an efficient way. Such message passing
can be implemented with convolution between features maps in the same layer,
and it is also integrated with feedforward propagation in neural networks.
Finally, a neural network implementation of end-to-end learning CRF-CNN is
provided. Its effectiveness is demonstrated through experiments on two
benchmark datasets.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 04:42:40 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Chu",
"Xiao",
""
],
[
"Ouyang",
"Wanli",
""
],
[
"Li",
"Hongsheng",
""
],
[
"Wang",
"Xiaogang",
""
]
] | TITLE: CRF-CNN: Modeling Structured Information in Human Pose Estimation
ABSTRACT: Deep convolutional neural networks (CNN) have achieved great success. On the
other hand, modeling structural information has been proved critical in many
vision problems. It is of great interest to integrate them effectively. In a
classical neural network, there is no message passing between neurons in the
same layer. In this paper, we propose a CRF-CNN framework which can
simultaneously model structural information in both output and hidden feature
layers in a probabilistic way, and it is applied to human pose estimation. A
message passing scheme is proposed, so that in various layers each body joint
receives messages from all the others in an efficient way. Such message passing
can be implemented with convolution between features maps in the same layer,
and it is also integrated with feedforward propagation in neural networks.
Finally, a neural network implementation of end-to-end learning CRF-CNN is
provided. Its effectiveness is demonstrated through experiments on two
benchmark datasets.
| no_new_dataset | 0.952838 |
1611.00472 | Ameya Prabhu | Ameya Prabhu, Aditya Joshi, Manish Shrivastava and Vasudeva Varma | Towards Sub-Word Level Compositions for Sentiment Analysis of
Hindi-English Code Mixed Text | Accepted paper at COLING 2016 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Sentiment analysis (SA) using code-mixed data from social media has several
applications in opinion mining ranging from customer satisfaction to social
campaign analysis in multilingual societies. Advances in this area are impeded
by the lack of a suitable annotated dataset. We introduce a Hindi-English
(Hi-En) code-mixed dataset for sentiment analysis and perform empirical
analysis comparing the suitability and performance of various state-of-the-art
SA methods in social media.
In this paper, we introduce learning sub-word level representations in LSTM
(Subword-LSTM) architecture instead of character-level or word-level
representations. This linguistic prior in our architecture enables us to learn
the information about sentiment value of important morphemes. This also seems
to work well in highly noisy text containing misspellings as shown in our
experiments which is demonstrated in morpheme-level feature maps learned by our
model. Also, we hypothesize that encoding this linguistic prior in the
Subword-LSTM architecture leads to the superior performance. Our system attains
accuracy 4-5% greater than traditional approaches on our dataset, and also
outperforms the available system for sentiment analysis in Hi-En code-mixed
text by 18%.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 05:23:53 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Prabhu",
"Ameya",
""
],
[
"Joshi",
"Aditya",
""
],
[
"Shrivastava",
"Manish",
""
],
[
"Varma",
"Vasudeva",
""
]
] | TITLE: Towards Sub-Word Level Compositions for Sentiment Analysis of
Hindi-English Code Mixed Text
ABSTRACT: Sentiment analysis (SA) using code-mixed data from social media has several
applications in opinion mining ranging from customer satisfaction to social
campaign analysis in multilingual societies. Advances in this area are impeded
by the lack of a suitable annotated dataset. We introduce a Hindi-English
(Hi-En) code-mixed dataset for sentiment analysis and perform empirical
analysis comparing the suitability and performance of various state-of-the-art
SA methods in social media.
In this paper, we introduce learning sub-word level representations in LSTM
(Subword-LSTM) architecture instead of character-level or word-level
representations. This linguistic prior in our architecture enables us to learn
the information about sentiment value of important morphemes. This also seems
to work well in highly noisy text containing misspellings as shown in our
experiments which is demonstrated in morpheme-level feature maps learned by our
model. Also, we hypothesize that encoding this linguistic prior in the
Subword-LSTM architecture leads to the superior performance. Our system attains
accuracy 4-5% greater than traditional approaches on our dataset, and also
outperforms the available system for sentiment analysis in Hi-En code-mixed
text by 18%.
| new_dataset | 0.971047 |
1611.00549 | Oliver Cliff | Oliver M. Cliff and Mikhail Prokopenko and Robert Fitch | Inferring Coupling of Distributed Dynamical Systems via Transfer Entropy | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we are interested in structure learning for a set of spatially
distributed dynamical systems, where individual subsystems are coupled via
latent variables and observed through a filter. We represent this model as a
directed acyclic graph (DAG) that characterises the unidirectional coupling
between subsystems. Standard approaches to structure learning are not
applicable in this framework due to the hidden variables, however we can
exploit the properties of certain dynamical systems to formulate exact methods
based on state space reconstruction. We approach the problem by using
reconstruction theorems to analytically derive a tractable expression for the
KL-divergence of a candidate DAG from the observed dataset. We show this
measure can be decomposed as a function of two information-theoretic measures,
transfer entropy and stochastic interaction. We then present two mathematically
robust scoring functions based on transfer entropy and statistical independence
tests. These results support the previously held conjecture that transfer
entropy can be used to infer effective connectivity in complex networks.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 11:23:54 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Cliff",
"Oliver M.",
""
],
[
"Prokopenko",
"Mikhail",
""
],
[
"Fitch",
"Robert",
""
]
] | TITLE: Inferring Coupling of Distributed Dynamical Systems via Transfer Entropy
ABSTRACT: In this work, we are interested in structure learning for a set of spatially
distributed dynamical systems, where individual subsystems are coupled via
latent variables and observed through a filter. We represent this model as a
directed acyclic graph (DAG) that characterises the unidirectional coupling
between subsystems. Standard approaches to structure learning are not
applicable in this framework due to the hidden variables, however we can
exploit the properties of certain dynamical systems to formulate exact methods
based on state space reconstruction. We approach the problem by using
reconstruction theorems to analytically derive a tractable expression for the
KL-divergence of a candidate DAG from the observed dataset. We show this
measure can be decomposed as a function of two information-theoretic measures,
transfer entropy and stochastic interaction. We then present two mathematically
robust scoring functions based on transfer entropy and statistical independence
tests. These results support the previously held conjecture that transfer
entropy can be used to infer effective connectivity in complex networks.
| no_new_dataset | 0.943348 |
1611.00714 | Alexander Jung | Alexander Jung and Alfred O. Hero III and Alexandru Mara and Sabeur
Aridhi | Scalable Semi-Supervised Learning over Networks using Nonsmooth Convex
Optimization | null | null | null | null | cs.LG cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a scalable method for semi-supervised (transductive) learning from
massive network-structured datasets. Our approach to semi-supervised learning
is based on representing the underlying hypothesis as a graph signal with small
total variation. Requiring a small total variation of the graph signal
representing the underlying hypothesis corresponds to the central smoothness
assumption that forms the basis for semi-supervised learning, i.e., input
points forming clusters have similar output values or labels. We formulate the
learning problem as a nonsmooth convex optimization problem which we solve by
appealing to Nesterovs optimal first-order method for nonsmooth optimization.
We also provide a message passing formulation of the learning method which
allows for a highly scalable implementation in big data frameworks.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2016 18:27:53 GMT"
}
] | 2016-11-03T00:00:00 | [
[
"Jung",
"Alexander",
""
],
[
"Hero",
"Alfred O.",
"III"
],
[
"Mara",
"Alexandru",
""
],
[
"Aridhi",
"Sabeur",
""
]
] | TITLE: Scalable Semi-Supervised Learning over Networks using Nonsmooth Convex
Optimization
ABSTRACT: We propose a scalable method for semi-supervised (transductive) learning from
massive network-structured datasets. Our approach to semi-supervised learning
is based on representing the underlying hypothesis as a graph signal with small
total variation. Requiring a small total variation of the graph signal
representing the underlying hypothesis corresponds to the central smoothness
assumption that forms the basis for semi-supervised learning, i.e., input
points forming clusters have similar output values or labels. We formulate the
learning problem as a nonsmooth convex optimization problem which we solve by
appealing to Nesterovs optimal first-order method for nonsmooth optimization.
We also provide a message passing formulation of the learning method which
allows for a highly scalable implementation in big data frameworks.
| no_new_dataset | 0.950457 |
1511.06049 | Deyu Meng | Deyu Meng and Qian Zhao and Lu Jiang | What Objective Does Self-paced Learning Indeed Optimize? | 25 pages, 1 figures | null | null | null | cs.LG cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | Self-paced learning (SPL) is a recently raised methodology designed through
simulating the learning principle of humans/animals. A variety of SPL
realization schemes have been designed for different computer vision and
pattern recognition tasks, and empirically substantiated to be effective in
these applications. However, the investigation on its theoretical insight is
still a blank. To this issue, this study attempts to provide some new
theoretical understanding under the SPL scheme. Specifically, we prove that the
solving strategy on SPL accords with a majorization minimization algorithm
implemented on a latent objective function. Furthermore, we find that the loss
function contained in this latent objective has a similar configuration with
non-convex regularized penalty (NSPR) known in statistics and machine learning.
Such connection inspires us discovering more intrinsic relationship between SPL
regimes and NSPR forms, like SCAD, LOG and EXP. The robustness insight under
SPL can then be finely explained. We also analyze the capability of SPL on its
easy loss prior embedding property, and provide an insightful interpretation to
the effectiveness mechanism under previous SPL variations. Besides, we design a
group-partial-order loss prior, which is especially useful to weakly labeled
large-scale data processing tasks. Through applying SPL with this loss prior to
the FCVID dataset, which is currently one of the biggest manually annotated
video dataset, our method achieves state-of-the-art performance beyond previous
methods, which further helps supports the proposed theoretical arguments.
| [
{
"version": "v1",
"created": "Thu, 19 Nov 2015 02:55:18 GMT"
},
{
"version": "v2",
"created": "Tue, 1 Nov 2016 13:59:27 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Meng",
"Deyu",
""
],
[
"Zhao",
"Qian",
""
],
[
"Jiang",
"Lu",
""
]
] | TITLE: What Objective Does Self-paced Learning Indeed Optimize?
ABSTRACT: Self-paced learning (SPL) is a recently raised methodology designed through
simulating the learning principle of humans/animals. A variety of SPL
realization schemes have been designed for different computer vision and
pattern recognition tasks, and empirically substantiated to be effective in
these applications. However, the investigation on its theoretical insight is
still a blank. To this issue, this study attempts to provide some new
theoretical understanding under the SPL scheme. Specifically, we prove that the
solving strategy on SPL accords with a majorization minimization algorithm
implemented on a latent objective function. Furthermore, we find that the loss
function contained in this latent objective has a similar configuration with
non-convex regularized penalty (NSPR) known in statistics and machine learning.
Such connection inspires us discovering more intrinsic relationship between SPL
regimes and NSPR forms, like SCAD, LOG and EXP. The robustness insight under
SPL can then be finely explained. We also analyze the capability of SPL on its
easy loss prior embedding property, and provide an insightful interpretation to
the effectiveness mechanism under previous SPL variations. Besides, we design a
group-partial-order loss prior, which is especially useful to weakly labeled
large-scale data processing tasks. Through applying SPL with this loss prior to
the FCVID dataset, which is currently one of the biggest manually annotated
video dataset, our method achieves state-of-the-art performance beyond previous
methods, which further helps supports the proposed theoretical arguments.
| no_new_dataset | 0.942401 |
1609.03426 | Sayantan Dasgupta | Sayantan Dasgupta | Multi-Label Learning with Provable Guarantee | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Here we study the problem of learning labels for large text corpora where
each text can be assigned a variable number of labels. The problem might seem
trivial when the label dimensionality is small and can be easily solved using a
series of one-vs-all classifiers. However, as the label dimensionality
increases to several thousand, the parameter space becomes extremely large, and
it is no longer possible to use the one-vs-all technique. Here we propose a
model based on the factorization of higher order moments of the words in the
corpora, as well as the cross moment between the labels and the words for
multi-label prediction. Our model provides guaranteed convergence bounds on the
estimated parameters. Further, our model takes only three passes through the
training dataset to extract the parameters, resulting in a highly scalable
algorithm that can train on GB's of data consisting of millions of documents
with hundreds of thousands of labels using a nominal resource of a single
processor with 16GB RAM. Our model achieves 10x-15x order of speed-up on
large-scale datasets while producing competitive performance in comparison with
existing benchmark algorithms.
| [
{
"version": "v1",
"created": "Mon, 12 Sep 2016 14:38:08 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Sep 2016 23:26:50 GMT"
},
{
"version": "v3",
"created": "Sun, 18 Sep 2016 14:57:20 GMT"
},
{
"version": "v4",
"created": "Tue, 1 Nov 2016 16:21:54 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Dasgupta",
"Sayantan",
""
]
] | TITLE: Multi-Label Learning with Provable Guarantee
ABSTRACT: Here we study the problem of learning labels for large text corpora where
each text can be assigned a variable number of labels. The problem might seem
trivial when the label dimensionality is small and can be easily solved using a
series of one-vs-all classifiers. However, as the label dimensionality
increases to several thousand, the parameter space becomes extremely large, and
it is no longer possible to use the one-vs-all technique. Here we propose a
model based on the factorization of higher order moments of the words in the
corpora, as well as the cross moment between the labels and the words for
multi-label prediction. Our model provides guaranteed convergence bounds on the
estimated parameters. Further, our model takes only three passes through the
training dataset to extract the parameters, resulting in a highly scalable
algorithm that can train on GB's of data consisting of millions of documents
with hundreds of thousands of labels using a nominal resource of a single
processor with 16GB RAM. Our model achieves 10x-15x order of speed-up on
large-scale datasets while producing competitive performance in comparison with
existing benchmark algorithms.
| no_new_dataset | 0.944689 |
1609.04855 | Kenji Hata | Kenji Hata, Ranjay Krishna, Li Fei-Fei, Michael S. Bernstein | A Glimpse Far into the Future: Understanding Long-term Crowd Worker
Quality | 10 pages, 11 figures, accepted CSCW 2017 | null | 10.1145/2998181.2998248 | null | cs.HC cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Microtask crowdsourcing is increasingly critical to the creation of extremely
large datasets. As a result, crowd workers spend weeks or months repeating the
exact same tasks, making it necessary to understand their behavior over these
long periods of time. We utilize three large, longitudinal datasets of nine
million annotations collected from Amazon Mechanical Turk to examine claims
that workers fatigue or satisfice over these long periods, producing lower
quality work. We find that, contrary to these claims, workers are extremely
stable in their quality over the entire period. To understand whether workers
set their quality based on the task's requirements for acceptance, we then
perform an experiment where we vary the required quality for a large
crowdsourcing task. Workers did not adjust their quality based on the
acceptance threshold: workers who were above the threshold continued working at
their usual quality level, and workers below the threshold self-selected
themselves out of the task. Capitalizing on this consistency, we demonstrate
that it is possible to predict workers' long-term quality using just a glimpse
of their quality on the first five tasks.
| [
{
"version": "v1",
"created": "Thu, 15 Sep 2016 20:47:51 GMT"
},
{
"version": "v2",
"created": "Tue, 1 Nov 2016 17:34:10 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Hata",
"Kenji",
""
],
[
"Krishna",
"Ranjay",
""
],
[
"Fei-Fei",
"Li",
""
],
[
"Bernstein",
"Michael S.",
""
]
] | TITLE: A Glimpse Far into the Future: Understanding Long-term Crowd Worker
Quality
ABSTRACT: Microtask crowdsourcing is increasingly critical to the creation of extremely
large datasets. As a result, crowd workers spend weeks or months repeating the
exact same tasks, making it necessary to understand their behavior over these
long periods of time. We utilize three large, longitudinal datasets of nine
million annotations collected from Amazon Mechanical Turk to examine claims
that workers fatigue or satisfice over these long periods, producing lower
quality work. We find that, contrary to these claims, workers are extremely
stable in their quality over the entire period. To understand whether workers
set their quality based on the task's requirements for acceptance, we then
perform an experiment where we vary the required quality for a large
crowdsourcing task. Workers did not adjust their quality based on the
acceptance threshold: workers who were above the threshold continued working at
their usual quality level, and workers below the threshold self-selected
themselves out of the task. Capitalizing on this consistency, we demonstrate
that it is possible to predict workers' long-term quality using just a glimpse
of their quality on the first five tasks.
| no_new_dataset | 0.924005 |
1611.00129 | Jiecao Chen | Jiecao Chen, Huy L. Nguyen, Qin Zhang | Submodular Maximization over Sliding Windows | 13 pages | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we study the extraction of representative elements in the data
stream model in the form of submodular maximization. Different from the
previous work on streaming submodular maximization, we are interested only in
the recent data, and study the maximization problem over sliding windows. We
provide a general reduction from the sliding window model to the standard
streaming model, and thus our approach works for general constraints as long as
there is a corresponding streaming algorithm in the standard streaming model.
As a consequence, we obtain the first algorithms in the sliding window model
for maximizing a monotone/non-monotone submodular function under cardinality
and matroid constraints. We also propose several heuristics and show their
efficiency in real-world datasets.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 05:06:37 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Chen",
"Jiecao",
""
],
[
"Nguyen",
"Huy L.",
""
],
[
"Zhang",
"Qin",
""
]
] | TITLE: Submodular Maximization over Sliding Windows
ABSTRACT: In this paper we study the extraction of representative elements in the data
stream model in the form of submodular maximization. Different from the
previous work on streaming submodular maximization, we are interested only in
the recent data, and study the maximization problem over sliding windows. We
provide a general reduction from the sliding window model to the standard
streaming model, and thus our approach works for general constraints as long as
there is a corresponding streaming algorithm in the standard streaming model.
As a consequence, we obtain the first algorithms in the sliding window model
for maximizing a monotone/non-monotone submodular function under cardinality
and matroid constraints. We also propose several heuristics and show their
efficiency in real-world datasets.
| no_new_dataset | 0.951684 |
1611.00144 | Yanru Qu | Yanru Qu, Han Cai, Kan Ren, Weinan Zhang, Yong Yu, Ying Wen, Jun Wang | Product-based Neural Networks for User Response Prediction | 6 pages, 5 figures, ICDM2016 | null | null | null | cs.LG cs.IR | http://creativecommons.org/licenses/by/4.0/ | Predicting user responses, such as clicks and conversions, is of great
importance and has found its usage in many Web applications including
recommender systems, web search and online advertising. The data in those
applications is mostly categorical and contains multiple fields; a typical
representation is to transform it into a high-dimensional sparse binary feature
representation via one-hot encoding. Facing with the extreme sparsity,
traditional models may limit their capacity of mining shallow patterns from the
data, i.e. low-order feature combinations. Deep models like deep neural
networks, on the other hand, cannot be directly applied for the
high-dimensional input because of the huge feature space. In this paper, we
propose a Product-based Neural Networks (PNN) with an embedding layer to learn
a distributed representation of the categorical data, a product layer to
capture interactive patterns between inter-field categories, and further fully
connected layers to explore high-order feature interactions. Our experimental
results on two large-scale real-world ad click datasets demonstrate that PNNs
consistently outperform the state-of-the-art models on various metrics.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 07:10:22 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Qu",
"Yanru",
""
],
[
"Cai",
"Han",
""
],
[
"Ren",
"Kan",
""
],
[
"Zhang",
"Weinan",
""
],
[
"Yu",
"Yong",
""
],
[
"Wen",
"Ying",
""
],
[
"Wang",
"Jun",
""
]
] | TITLE: Product-based Neural Networks for User Response Prediction
ABSTRACT: Predicting user responses, such as clicks and conversions, is of great
importance and has found its usage in many Web applications including
recommender systems, web search and online advertising. The data in those
applications is mostly categorical and contains multiple fields; a typical
representation is to transform it into a high-dimensional sparse binary feature
representation via one-hot encoding. Facing with the extreme sparsity,
traditional models may limit their capacity of mining shallow patterns from the
data, i.e. low-order feature combinations. Deep models like deep neural
networks, on the other hand, cannot be directly applied for the
high-dimensional input because of the huge feature space. In this paper, we
propose a Product-based Neural Networks (PNN) with an embedding layer to learn
a distributed representation of the categorical data, a product layer to
capture interactive patterns between inter-field categories, and further fully
connected layers to explore high-order feature interactions. Our experimental
results on two large-scale real-world ad click datasets demonstrate that PNNs
consistently outperform the state-of-the-art models on various metrics.
| no_new_dataset | 0.948202 |
1611.00218 | Yashas Annadani | Yashas Annadani, D L Rakshith, Soma Biswas | Sliding Dictionary Based Sparse Representation For Action Recognition | 7 Pages | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The task of action recognition has been in the forefront of research, given
its applications in gaming, surveillance and health care. In this work, we
propose a simple, yet very effective approach which works seamlessly for both
offline and online action recognition using the skeletal joints. We construct a
sliding dictionary which has the training data along with their time stamps.
This is used to compute the sparse coefficients of the input action sequence
which is divided into overlapping windows and each window gives a probability
score for each action class. In addition, we compute another simple feature,
which calibrates each of the action sequences to the training sequences, and
models the deviation of the action from the each of the training data. Finally,
a score level fusion of the two heterogeneous but complementary features for
each window is obtained and the scores for the available windows are
successively combined to give the confidence scores of each action class. This
way of combining the scores makes the approach suitable for scenarios where
only part of the sequence is available. Extensive experimental evaluation on
three publicly available datasets shows the effectiveness of the proposed
approach for both offline and online action recognition tasks.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 13:29:38 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Annadani",
"Yashas",
""
],
[
"Rakshith",
"D L",
""
],
[
"Biswas",
"Soma",
""
]
] | TITLE: Sliding Dictionary Based Sparse Representation For Action Recognition
ABSTRACT: The task of action recognition has been in the forefront of research, given
its applications in gaming, surveillance and health care. In this work, we
propose a simple, yet very effective approach which works seamlessly for both
offline and online action recognition using the skeletal joints. We construct a
sliding dictionary which has the training data along with their time stamps.
This is used to compute the sparse coefficients of the input action sequence
which is divided into overlapping windows and each window gives a probability
score for each action class. In addition, we compute another simple feature,
which calibrates each of the action sequences to the training sequences, and
models the deviation of the action from the each of the training data. Finally,
a score level fusion of the two heterogeneous but complementary features for
each window is obtained and the scores for the available windows are
successively combined to give the confidence scores of each action class. This
way of combining the scores makes the approach suitable for scenarios where
only part of the sequence is available. Extensive experimental evaluation on
three publicly available datasets shows the effectiveness of the proposed
approach for both offline and online action recognition tasks.
| no_new_dataset | 0.943504 |
1611.00291 | Vikram Krishnamurthy | Vikram Krishnamurthy and Anup Aprem and Sujay Bhatt | Opportunistic Advertisement Scheduling in Live Social Media: A Multiple
Stopping Time POMDP Approach | null | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Live online social broadcasting services like YouTube Live and Twitch have
steadily gained popularity due to improved bandwidth, ease of generating
content and the ability to earn revenue on the generated content. In contrast
to traditional cable television, revenue in online services is generated solely
through advertisements, and depends on the number of clicks generated. Channel
owners aim to opportunistically schedule advertisements so as to generate
maximum revenue. This paper considers the problem of optimal scheduling of
advertisements in live online social media. The problem is formulated as a
multiple stopping problem and is addressed in a partially observed Markov
decision process (POMDP) framework. Structural results are provided on the
optimal advertisement scheduling policy. By exploiting the structure of the
optimal policy, best linear thresholds are computed using stochastic
approximation. The proposed model and framework are validated on real datasets,
and the following observations are made: (i) The policy obtained by the
multiple stopping problem can be used to detect changes in ground truth from
online search data (ii) Numerical results show a significant improvement in the
expected revenue by opportunistically scheduling the advertisements. The
revenue can be improved by $20-30\%$ in comparison to currently employed
periodic scheduling.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2016 16:48:10 GMT"
}
] | 2016-11-02T00:00:00 | [
[
"Krishnamurthy",
"Vikram",
""
],
[
"Aprem",
"Anup",
""
],
[
"Bhatt",
"Sujay",
""
]
] | TITLE: Opportunistic Advertisement Scheduling in Live Social Media: A Multiple
Stopping Time POMDP Approach
ABSTRACT: Live online social broadcasting services like YouTube Live and Twitch have
steadily gained popularity due to improved bandwidth, ease of generating
content and the ability to earn revenue on the generated content. In contrast
to traditional cable television, revenue in online services is generated solely
through advertisements, and depends on the number of clicks generated. Channel
owners aim to opportunistically schedule advertisements so as to generate
maximum revenue. This paper considers the problem of optimal scheduling of
advertisements in live online social media. The problem is formulated as a
multiple stopping problem and is addressed in a partially observed Markov
decision process (POMDP) framework. Structural results are provided on the
optimal advertisement scheduling policy. By exploiting the structure of the
optimal policy, best linear thresholds are computed using stochastic
approximation. The proposed model and framework are validated on real datasets,
and the following observations are made: (i) The policy obtained by the
multiple stopping problem can be used to detect changes in ground truth from
online search data (ii) Numerical results show a significant improvement in the
expected revenue by opportunistically scheduling the advertisements. The
revenue can be improved by $20-30\%$ in comparison to currently employed
periodic scheduling.
| no_new_dataset | 0.950273 |
1412.3708 | Marc Goessling | Marc Goessling and Yali Amit | Compact Compositional Models | null | null | null | null | cs.CV cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning compact and interpretable representations is a very natural task,
which has not been solved satisfactorily even for simple binary datasets. In
this paper, we review various ways of composing experts for binary data and
argue that competitive forms of interaction are best suited to learn
low-dimensional representations. We propose a new composition rule that
discourages experts from focusing on similar structures and that penalizes
opposing votes strongly so that abstaining from voting becomes more attractive.
We also introduce a novel sequential initialization procedure, which is based
on a process of oversimplification and correction. Experiments show that with
our approach very intuitive models can be learned.
| [
{
"version": "v1",
"created": "Thu, 11 Dec 2014 16:19:56 GMT"
},
{
"version": "v2",
"created": "Wed, 25 Feb 2015 19:23:27 GMT"
},
{
"version": "v3",
"created": "Wed, 8 Apr 2015 22:02:42 GMT"
},
{
"version": "v4",
"created": "Sat, 29 Oct 2016 22:49:39 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Goessling",
"Marc",
""
],
[
"Amit",
"Yali",
""
]
] | TITLE: Compact Compositional Models
ABSTRACT: Learning compact and interpretable representations is a very natural task,
which has not been solved satisfactorily even for simple binary datasets. In
this paper, we review various ways of composing experts for binary data and
argue that competitive forms of interaction are best suited to learn
low-dimensional representations. We propose a new composition rule that
discourages experts from focusing on similar structures and that penalizes
opposing votes strongly so that abstaining from voting becomes more attractive.
We also introduce a novel sequential initialization procedure, which is based
on a process of oversimplification and correction. Experiments show that with
our approach very intuitive models can be learned.
| no_new_dataset | 0.948202 |
1510.09083 | Hanjiang Lai | Hanjiang Lai, Shengtao Xiao, Yan Pan, Zhen Cui, Jiashi Feng, Chunyan
Xu, Jian Yin and Shuicheng Yan | Deep Recurrent Regression for Facial Landmark Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel end-to-end deep architecture for face landmark detection,
based on a deep convolutional and deconvolutional network followed by carefully
designed recurrent network structures. The pipeline of this architecture
consists of three parts. Through the first part, we encode an input face image
to resolution-preserved deconvolutional feature maps via a deep network with
stacked convolutional and deconvolutional layers. Then, in the second part, we
estimate the initial coordinates of the facial key points by an additional
convolutional layer on top of these deconvolutional feature maps. In the last
part, by using the deconvolutional feature maps and the initial facial key
points as input, we refine the coordinates of the facial key points by a
recurrent network that consists of multiple Long-Short Term Memory (LSTM)
components. Extensive evaluations on several benchmark datasets show that the
proposed deep architecture has superior performance against the
state-of-the-art methods.
| [
{
"version": "v1",
"created": "Fri, 30 Oct 2015 13:34:18 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Nov 2015 01:54:11 GMT"
},
{
"version": "v3",
"created": "Mon, 31 Oct 2016 03:29:54 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Lai",
"Hanjiang",
""
],
[
"Xiao",
"Shengtao",
""
],
[
"Pan",
"Yan",
""
],
[
"Cui",
"Zhen",
""
],
[
"Feng",
"Jiashi",
""
],
[
"Xu",
"Chunyan",
""
],
[
"Yin",
"Jian",
""
],
[
"Yan",
"Shuicheng",
""
]
] | TITLE: Deep Recurrent Regression for Facial Landmark Detection
ABSTRACT: We propose a novel end-to-end deep architecture for face landmark detection,
based on a deep convolutional and deconvolutional network followed by carefully
designed recurrent network structures. The pipeline of this architecture
consists of three parts. Through the first part, we encode an input face image
to resolution-preserved deconvolutional feature maps via a deep network with
stacked convolutional and deconvolutional layers. Then, in the second part, we
estimate the initial coordinates of the facial key points by an additional
convolutional layer on top of these deconvolutional feature maps. In the last
part, by using the deconvolutional feature maps and the initial facial key
points as input, we refine the coordinates of the facial key points by a
recurrent network that consists of multiple Long-Short Term Memory (LSTM)
components. Extensive evaluations on several benchmark datasets show that the
proposed deep architecture has superior performance against the
state-of-the-art methods.
| no_new_dataset | 0.943867 |
1511.05933 | Sayantan Dasgupta | Sayantan Dasgupta | Seeding K-Means using Method of Moments | Paper contained an error in Equation 5 and 7 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | K-means is one of the most widely used algorithms for clustering in Data
Mining applications, which attempts to minimize the sum of the square of the
Euclidean distance of the points in the clusters from the respective means of
the clusters. However, K-means suffers from local minima problem and is not
guaranteed to converge to the optimal cost. K-means++ tries to address the
problem by seeding the means using a distance-based sampling scheme. However,
seeding the means in K-means++ needs $O\left(K\right)$ sequential passes
through the entire dataset, and this can be very costly for large datasets.
Here we propose a method of seeding the initial means based on factorizations
of higher order moments for bounded data. Our method takes $O\left(1\right)$
passes through the entire dataset to extract the initial set of means, and its
final cost can be proven to be within $O(\sqrt{K})$ of the optimal cost. We
demonstrate the performance of our algorithm in comparison with the existing
algorithms on various benchmark datasets.
| [
{
"version": "v1",
"created": "Wed, 18 Nov 2015 20:26:42 GMT"
},
{
"version": "v2",
"created": "Sat, 21 Nov 2015 21:54:01 GMT"
},
{
"version": "v3",
"created": "Thu, 4 Feb 2016 10:21:55 GMT"
},
{
"version": "v4",
"created": "Thu, 3 Mar 2016 17:40:02 GMT"
},
{
"version": "v5",
"created": "Thu, 21 Apr 2016 21:50:39 GMT"
},
{
"version": "v6",
"created": "Fri, 3 Jun 2016 17:50:10 GMT"
},
{
"version": "v7",
"created": "Mon, 12 Sep 2016 22:33:06 GMT"
},
{
"version": "v8",
"created": "Mon, 31 Oct 2016 15:59:13 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Dasgupta",
"Sayantan",
""
]
] | TITLE: Seeding K-Means using Method of Moments
ABSTRACT: K-means is one of the most widely used algorithms for clustering in Data
Mining applications, which attempts to minimize the sum of the square of the
Euclidean distance of the points in the clusters from the respective means of
the clusters. However, K-means suffers from local minima problem and is not
guaranteed to converge to the optimal cost. K-means++ tries to address the
problem by seeding the means using a distance-based sampling scheme. However,
seeding the means in K-means++ needs $O\left(K\right)$ sequential passes
through the entire dataset, and this can be very costly for large datasets.
Here we propose a method of seeding the initial means based on factorizations
of higher order moments for bounded data. Our method takes $O\left(1\right)$
passes through the entire dataset to extract the initial set of means, and its
final cost can be proven to be within $O(\sqrt{K})$ of the optimal cost. We
demonstrate the performance of our algorithm in comparison with the existing
algorithms on various benchmark datasets.
| no_new_dataset | 0.946941 |
1512.02363 | Christian Forster | Christian Forster, Luca Carlone, Frank Dellaert, Davide Scaramuzza | On-Manifold Preintegration for Real-Time Visual-Inertial Odometry | 20 pages, 24 figures, accepted for publication in IEEE Transactions
on Robotics (TRO) 2016 | null | 10.1109/TRO.2016.2597321 | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current approaches for visual-inertial odometry (VIO) are able to attain
highly accurate state estimation via nonlinear optimization. However, real-time
optimization quickly becomes infeasible as the trajectory grows over time, this
problem is further emphasized by the fact that inertial measurements come at
high rate, hence leading to fast growth of the number of variables in the
optimization. In this paper, we address this issue by preintegrating inertial
measurements between selected keyframes into single relative motion
constraints. Our first contribution is a \emph{preintegration theory} that
properly addresses the manifold structure of the rotation group. We formally
discuss the generative measurement model as well as the nature of the rotation
noise and derive the expression for the \emph{maximum a posteriori} state
estimator. Our theoretical development enables the computation of all necessary
Jacobians for the optimization and a-posteriori bias correction in analytic
form. The second contribution is to show that the preintegrated IMU model can
be seamlessly integrated into a visual-inertial pipeline under the unifying
framework of factor graphs. This enables the application of
incremental-smoothing algorithms and the use of a \emph{structureless} model
for visual measurements, which avoids optimizing over the 3D points, further
accelerating the computation. We perform an extensive evaluation of our
monocular \VIO pipeline on real and simulated datasets. The results confirm
that our modelling effort leads to accurate state estimation in real-time,
outperforming state-of-the-art approaches.
| [
{
"version": "v1",
"created": "Tue, 8 Dec 2015 08:26:25 GMT"
},
{
"version": "v2",
"created": "Tue, 19 Jul 2016 07:24:26 GMT"
},
{
"version": "v3",
"created": "Sun, 30 Oct 2016 10:43:58 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Forster",
"Christian",
""
],
[
"Carlone",
"Luca",
""
],
[
"Dellaert",
"Frank",
""
],
[
"Scaramuzza",
"Davide",
""
]
] | TITLE: On-Manifold Preintegration for Real-Time Visual-Inertial Odometry
ABSTRACT: Current approaches for visual-inertial odometry (VIO) are able to attain
highly accurate state estimation via nonlinear optimization. However, real-time
optimization quickly becomes infeasible as the trajectory grows over time, this
problem is further emphasized by the fact that inertial measurements come at
high rate, hence leading to fast growth of the number of variables in the
optimization. In this paper, we address this issue by preintegrating inertial
measurements between selected keyframes into single relative motion
constraints. Our first contribution is a \emph{preintegration theory} that
properly addresses the manifold structure of the rotation group. We formally
discuss the generative measurement model as well as the nature of the rotation
noise and derive the expression for the \emph{maximum a posteriori} state
estimator. Our theoretical development enables the computation of all necessary
Jacobians for the optimization and a-posteriori bias correction in analytic
form. The second contribution is to show that the preintegrated IMU model can
be seamlessly integrated into a visual-inertial pipeline under the unifying
framework of factor graphs. This enables the application of
incremental-smoothing algorithms and the use of a \emph{structureless} model
for visual measurements, which avoids optimizing over the 3D points, further
accelerating the computation. We perform an extensive evaluation of our
monocular \VIO pipeline on real and simulated datasets. The results confirm
that our modelling effort leads to accurate state estimation in real-time,
outperforming state-of-the-art approaches.
| no_new_dataset | 0.947721 |
1604.08672 | Shufeng Xiong | Shufeng Xiong, Yue Zhang, Donghong Ji, Yinxia Lou | Distance Metric Learning for Aspect Phrase Grouping | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aspect phrase grouping is an important task in aspect-level sentiment
analysis. It is a challenging problem due to polysemy and context dependency.
We propose an Attention-based Deep Distance Metric Learning (ADDML) method, by
considering aspect phrase representation as well as context representation.
First, leveraging the characteristics of the review text, we automatically
generate aspect phrase sample pairs for distant supervision. Second, we feed
word embeddings of aspect phrases and their contexts into an attention-based
neural network to learn feature representation of contexts. Both aspect phrase
embedding and context embedding are used to learn a deep feature subspace for
measure the distances between aspect phrases for K-means clustering.
Experiments on four review datasets show that the proposed method outperforms
state-of-the-art strong baseline methods.
| [
{
"version": "v1",
"created": "Fri, 29 Apr 2016 02:44:02 GMT"
},
{
"version": "v2",
"created": "Sun, 30 Oct 2016 02:09:15 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Xiong",
"Shufeng",
""
],
[
"Zhang",
"Yue",
""
],
[
"Ji",
"Donghong",
""
],
[
"Lou",
"Yinxia",
""
]
] | TITLE: Distance Metric Learning for Aspect Phrase Grouping
ABSTRACT: Aspect phrase grouping is an important task in aspect-level sentiment
analysis. It is a challenging problem due to polysemy and context dependency.
We propose an Attention-based Deep Distance Metric Learning (ADDML) method, by
considering aspect phrase representation as well as context representation.
First, leveraging the characteristics of the review text, we automatically
generate aspect phrase sample pairs for distant supervision. Second, we feed
word embeddings of aspect phrases and their contexts into an attention-based
neural network to learn feature representation of contexts. Both aspect phrase
embedding and context embedding are used to learn a deep feature subspace for
measure the distances between aspect phrases for K-means clustering.
Experiments on four review datasets show that the proposed method outperforms
state-of-the-art strong baseline methods.
| no_new_dataset | 0.949012 |
1606.00487 | Sepehr Valipour | Sepehr Valipour, Mennatullah Siam, Martin Jagersand, Nilanjan Ray | Recurrent Fully Convolutional Networks for Video Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image segmentation is an important step in most visual tasks. While
convolutional neural networks have shown to perform well on single image
segmentation, to our knowledge, no study has been been done on leveraging
recurrent gated architectures for video segmentation. Accordingly, we propose a
novel method for online segmentation of video sequences that incorporates
temporal data. The network is built from fully convolutional element and
recurrent unit that works on a sliding window over the temporal data. We also
introduce a novel convolutional gated recurrent unit that preserves the spatial
information and reduces the parameters learned. Our method has the advantage
that it can work in an online fashion instead of operating over the whole input
batch of video frames. The network is tested on the change detection dataset,
and proved to have 5.5\% improvement in F-measure over a plain fully
convolutional network for per frame segmentation. It was also shown to have
improvement of 1.4\% for the F-measure compared to our baseline network that we
call FCN 12s.
| [
{
"version": "v1",
"created": "Wed, 1 Jun 2016 22:27:41 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Jun 2016 07:24:00 GMT"
},
{
"version": "v3",
"created": "Mon, 31 Oct 2016 00:05:49 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Valipour",
"Sepehr",
""
],
[
"Siam",
"Mennatullah",
""
],
[
"Jagersand",
"Martin",
""
],
[
"Ray",
"Nilanjan",
""
]
] | TITLE: Recurrent Fully Convolutional Networks for Video Segmentation
ABSTRACT: Image segmentation is an important step in most visual tasks. While
convolutional neural networks have shown to perform well on single image
segmentation, to our knowledge, no study has been been done on leveraging
recurrent gated architectures for video segmentation. Accordingly, we propose a
novel method for online segmentation of video sequences that incorporates
temporal data. The network is built from fully convolutional element and
recurrent unit that works on a sliding window over the temporal data. We also
introduce a novel convolutional gated recurrent unit that preserves the spatial
information and reduces the parameters learned. Our method has the advantage
that it can work in an online fashion instead of operating over the whole input
batch of video frames. The network is tested on the change detection dataset,
and proved to have 5.5\% improvement in F-measure over a plain fully
convolutional network for per frame segmentation. It was also shown to have
improvement of 1.4\% for the F-measure compared to our baseline network that we
call FCN 12s.
| no_new_dataset | 0.950134 |
1606.03558 | Christopher Bongsoo Choy | Christopher B. Choy, JunYoung Gwak, Silvio Savarese, Manmohan
Chandraker | Universal Correspondence Network | To appear at NIPS 2016 as full oral presentation | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a deep learning framework for accurate visual correspondences and
demonstrate its effectiveness for both geometric and semantic matching,
spanning across rigid motions to intra-class shape or appearance variations. In
contrast to previous CNN-based approaches that optimize a surrogate patch
similarity objective, we use deep metric learning to directly learn a feature
space that preserves either geometric or semantic similarity. Our fully
convolutional architecture, along with a novel correspondence contrastive loss
allows faster training by effective reuse of computations, accurate gradient
computation through the use of thousands of examples per image pair and faster
testing with $O(n)$ feed forward passes for $n$ keypoints, instead of $O(n^2)$
for typical patch similarity methods. We propose a convolutional spatial
transformer to mimic patch normalization in traditional features like SIFT,
which is shown to dramatically boost accuracy for semantic correspondences
across intra-class shape variations. Extensive experiments on KITTI, PASCAL,
and CUB-2011 datasets demonstrate the significant advantages of our features
over prior works that use either hand-constructed or learned features.
| [
{
"version": "v1",
"created": "Sat, 11 Jun 2016 06:27:09 GMT"
},
{
"version": "v2",
"created": "Tue, 14 Jun 2016 23:16:13 GMT"
},
{
"version": "v3",
"created": "Mon, 31 Oct 2016 06:32:03 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Choy",
"Christopher B.",
""
],
[
"Gwak",
"JunYoung",
""
],
[
"Savarese",
"Silvio",
""
],
[
"Chandraker",
"Manmohan",
""
]
] | TITLE: Universal Correspondence Network
ABSTRACT: We present a deep learning framework for accurate visual correspondences and
demonstrate its effectiveness for both geometric and semantic matching,
spanning across rigid motions to intra-class shape or appearance variations. In
contrast to previous CNN-based approaches that optimize a surrogate patch
similarity objective, we use deep metric learning to directly learn a feature
space that preserves either geometric or semantic similarity. Our fully
convolutional architecture, along with a novel correspondence contrastive loss
allows faster training by effective reuse of computations, accurate gradient
computation through the use of thousands of examples per image pair and faster
testing with $O(n)$ feed forward passes for $n$ keypoints, instead of $O(n^2)$
for typical patch similarity methods. We propose a convolutional spatial
transformer to mimic patch normalization in traditional features like SIFT,
which is shown to dramatically boost accuracy for semantic correspondences
across intra-class shape variations. Extensive experiments on KITTI, PASCAL,
and CUB-2011 datasets demonstrate the significant advantages of our features
over prior works that use either hand-constructed or learned features.
| no_new_dataset | 0.946794 |
1607.02937 | Gabriel Gon\c{c}alves | Gabriel Resende Gon\c{c}alves, Sirlene Pio Gomes da Silva, David
Menotti, William Robson Schwartz | Benchmark for License Plate Character Segmentation | 32 pages, single column | J. Electron. Imaging. 25(5), 053034 (Oct 24, 2016) | 10.1117/1.JEI.25.5.053034 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatic License Plate Recognition (ALPR) has been the focus of many
researches in the past years. In general, ALPR is divided into the following
problems: detection of on-track vehicles, license plates detection, segmention
of license plate characters and optical character recognition (OCR). Even
though commercial solutions are available for controlled acquisition
conditions, e.g., the entrance of a parking lot, ALPR is still an open problem
when dealing with data acquired from uncontrolled environments, such as roads
and highways when relying only on imaging sensors. Due to the multiple
orientations and scales of the license plates captured by the camera, a very
challenging task of the ALPR is the License Plate Character Segmentation (LPCS)
step, which effectiveness is required to be (near) optimal to achieve a high
recognition rate by the OCR. To tackle the LPCS problem, this work proposes a
novel benchmark composed of a dataset designed to focus specifically on the
character segmentation step of the ALPR within an evaluation protocol.
Furthermore, we propose the Jaccard-Centroid coefficient, a new evaluation
measure more suitable than the Jaccard coefficient regarding the location of
the bounding box within the ground-truth annotation. The dataset is composed of
2,000 Brazilian license plates consisting of 14,000 alphanumeric symbols and
their corresponding bounding box annotations. We also present a new
straightforward approach to perform LPCS efficiently. Finally, we provide an
experimental evaluation for the dataset based on four LPCS approaches and
demonstrate the importance of character segmentation for achieving an accurate
OCR.
| [
{
"version": "v1",
"created": "Mon, 11 Jul 2016 13:32:19 GMT"
},
{
"version": "v2",
"created": "Mon, 31 Oct 2016 16:11:21 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Gonçalves",
"Gabriel Resende",
""
],
[
"da Silva",
"Sirlene Pio Gomes",
""
],
[
"Menotti",
"David",
""
],
[
"Schwartz",
"William Robson",
""
]
] | TITLE: Benchmark for License Plate Character Segmentation
ABSTRACT: Automatic License Plate Recognition (ALPR) has been the focus of many
researches in the past years. In general, ALPR is divided into the following
problems: detection of on-track vehicles, license plates detection, segmention
of license plate characters and optical character recognition (OCR). Even
though commercial solutions are available for controlled acquisition
conditions, e.g., the entrance of a parking lot, ALPR is still an open problem
when dealing with data acquired from uncontrolled environments, such as roads
and highways when relying only on imaging sensors. Due to the multiple
orientations and scales of the license plates captured by the camera, a very
challenging task of the ALPR is the License Plate Character Segmentation (LPCS)
step, which effectiveness is required to be (near) optimal to achieve a high
recognition rate by the OCR. To tackle the LPCS problem, this work proposes a
novel benchmark composed of a dataset designed to focus specifically on the
character segmentation step of the ALPR within an evaluation protocol.
Furthermore, we propose the Jaccard-Centroid coefficient, a new evaluation
measure more suitable than the Jaccard coefficient regarding the location of
the bounding box within the ground-truth annotation. The dataset is composed of
2,000 Brazilian license plates consisting of 14,000 alphanumeric symbols and
their corresponding bounding box annotations. We also present a new
straightforward approach to perform LPCS efficiently. Finally, we provide an
experimental evaluation for the dataset based on four LPCS approaches and
demonstrate the importance of character segmentation for achieving an accurate
OCR.
| new_dataset | 0.976333 |
1610.04834 | Mohsen Ghafoorian | Mohsen Ghafoorian, Nico Karssemeijer, Tom Heskes, Inge van Uden, Clara
Sanchez, Geert Litjens, Frank-Erik de Leeuw, Bram van Ginneken, Elena
Marchiori and Bram Platel | Location Sensitive Deep Convolutional Neural Networks for Segmentation
of White Matter Hyperintensities | 13 pages, 8 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The anatomical location of imaging features is of crucial importance for
accurate diagnosis in many medical tasks. Convolutional neural networks (CNN)
have had huge successes in computer vision, but they lack the natural ability
to incorporate the anatomical location in their decision making process,
hindering success in some medical image analysis tasks.
In this paper, to integrate the anatomical location information into the
network, we propose several deep CNN architectures that consider multi-scale
patches or take explicit location features while training. We apply and compare
the proposed architectures for segmentation of white matter hyperintensities in
brain MR images on a large dataset. As a result, we observe that the CNNs that
incorporate location information substantially outperform a conventional
segmentation method with hand-crafted features as well as CNNs that do not
integrate location information. On a test set of 46 scans, the best
configuration of our networks obtained a Dice score of 0.791, compared to 0.797
for an independent human observer. Performance levels of the machine and the
independent human observer were not statistically significantly different
(p-value=0.17).
| [
{
"version": "v1",
"created": "Sun, 16 Oct 2016 09:35:36 GMT"
},
{
"version": "v2",
"created": "Sat, 29 Oct 2016 15:10:46 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Ghafoorian",
"Mohsen",
""
],
[
"Karssemeijer",
"Nico",
""
],
[
"Heskes",
"Tom",
""
],
[
"van Uden",
"Inge",
""
],
[
"Sanchez",
"Clara",
""
],
[
"Litjens",
"Geert",
""
],
[
"de Leeuw",
"Frank-Erik",
""
],
[
"van Ginneken",
"Bram",
""
],
[
"Marchiori",
"Elena",
""
],
[
"Platel",
"Bram",
""
]
] | TITLE: Location Sensitive Deep Convolutional Neural Networks for Segmentation
of White Matter Hyperintensities
ABSTRACT: The anatomical location of imaging features is of crucial importance for
accurate diagnosis in many medical tasks. Convolutional neural networks (CNN)
have had huge successes in computer vision, but they lack the natural ability
to incorporate the anatomical location in their decision making process,
hindering success in some medical image analysis tasks.
In this paper, to integrate the anatomical location information into the
network, we propose several deep CNN architectures that consider multi-scale
patches or take explicit location features while training. We apply and compare
the proposed architectures for segmentation of white matter hyperintensities in
brain MR images on a large dataset. As a result, we observe that the CNNs that
incorporate location information substantially outperform a conventional
segmentation method with hand-crafted features as well as CNNs that do not
integrate location information. On a test set of 46 scans, the best
configuration of our networks obtained a Dice score of 0.791, compared to 0.797
for an independent human observer. Performance levels of the machine and the
independent human observer were not statistically significantly different
(p-value=0.17).
| no_new_dataset | 0.951818 |
1610.09451 | Evan Sparks | Evan R. Sparks, Shivaram Venkataraman, Tomer Kaftan, Michael J.
Franklin, Benjamin Recht | KeystoneML: Optimizing Pipelines for Large-Scale Advanced Analytics | null | null | null | null | cs.LG cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern advanced analytics applications make use of machine learning
techniques and contain multiple steps of domain-specific and general-purpose
processing with high resource requirements. We present KeystoneML, a system
that captures and optimizes the end-to-end large-scale machine learning
applications for high-throughput training in a distributed environment with a
high-level API. This approach offers increased ease of use and higher
performance over existing systems for large scale learning. We demonstrate the
effectiveness of KeystoneML in achieving high quality statistical accuracy and
scalable training using real world datasets in several domains. By optimizing
execution KeystoneML achieves up to 15x training throughput over unoptimized
execution on a real image classification application.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 04:21:24 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Sparks",
"Evan R.",
""
],
[
"Venkataraman",
"Shivaram",
""
],
[
"Kaftan",
"Tomer",
""
],
[
"Franklin",
"Michael J.",
""
],
[
"Recht",
"Benjamin",
""
]
] | TITLE: KeystoneML: Optimizing Pipelines for Large-Scale Advanced Analytics
ABSTRACT: Modern advanced analytics applications make use of machine learning
techniques and contain multiple steps of domain-specific and general-purpose
processing with high resource requirements. We present KeystoneML, a system
that captures and optimizes the end-to-end large-scale machine learning
applications for high-throughput training in a distributed environment with a
high-level API. This approach offers increased ease of use and higher
performance over existing systems for large scale learning. We demonstrate the
effectiveness of KeystoneML in achieving high quality statistical accuracy and
scalable training using real world datasets in several domains. By optimizing
execution KeystoneML achieves up to 15x training throughput over unoptimized
execution on a real image classification application.
| no_new_dataset | 0.948106 |
1610.09462 | Ye Liu | Ye Liu, Yuxuan Liang, Shuming Liu, David S. Rosenblum, and Yu Zheng | Predicting Urban Water Quality with Ubiquitous Data | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Urban water quality is of great importance to our daily lives. Prediction of
urban water quality help control water pollution and protect human health.
However, predicting the urban water quality is a challenging task since the
water quality varies in urban spaces non-linearly and depends on multiple
factors, such as meteorology, water usage patterns, and land uses. In this
work, we forecast the water quality of a station over the next few hours from a
data-driven perspective, using the water quality data and water hydraulic data
reported by existing monitor stations and a variety of data sources we observed
in the city, such as meteorology, pipe networks, structure of road networks,
and point of interests (POIs). First, we identify the influential factors that
affect the urban water quality via extensive experiments. Second, we present a
multi-task multi-view learning method to fuse those multiple datasets from
different domains into an unified learning model. We evaluate our method with
real-world datasets, and the extensive experiments verify the advantages of our
method over other baselines and demonstrate the effectiveness of our approach.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 06:04:14 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Liu",
"Ye",
""
],
[
"Liang",
"Yuxuan",
""
],
[
"Liu",
"Shuming",
""
],
[
"Rosenblum",
"David S.",
""
],
[
"Zheng",
"Yu",
""
]
] | TITLE: Predicting Urban Water Quality with Ubiquitous Data
ABSTRACT: Urban water quality is of great importance to our daily lives. Prediction of
urban water quality help control water pollution and protect human health.
However, predicting the urban water quality is a challenging task since the
water quality varies in urban spaces non-linearly and depends on multiple
factors, such as meteorology, water usage patterns, and land uses. In this
work, we forecast the water quality of a station over the next few hours from a
data-driven perspective, using the water quality data and water hydraulic data
reported by existing monitor stations and a variety of data sources we observed
in the city, such as meteorology, pipe networks, structure of road networks,
and point of interests (POIs). First, we identify the influential factors that
affect the urban water quality via extensive experiments. Second, we present a
multi-task multi-view learning method to fuse those multiple datasets from
different domains into an unified learning model. We evaluate our method with
real-world datasets, and the extensive experiments verify the advantages of our
method over other baselines and demonstrate the effectiveness of our approach.
| no_new_dataset | 0.947721 |
1610.09491 | Kiarash Shaloudegi | Kiarash Shaloudegi, Andr\'as Gy\"orgy, Csaba Szepesv\'ari, and Wilsun
Xu | SDP Relaxation with Randomized Rounding for Energy Disaggregation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We develop a scalable, computationally efficient method for the task of
energy disaggregation for home appliance monitoring. In this problem the goal
is to estimate the energy consumption of each appliance over time based on the
total energy-consumption signal of a household. The current state of the art is
to model the problem as inference in factorial HMMs, and use quadratic
programming to find an approximate solution to the resulting quadratic integer
program. Here we take a more principled approach, better suited to integer
programming problems, and find an approximate optimum by combining convex
semidefinite relaxations randomized rounding, as well as a scalable ADMM method
that exploits the special structure of the resulting semidefinite program.
Simulation results both in synthetic and real-world datasets demonstrate the
superiority of our method.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 11:48:28 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Shaloudegi",
"Kiarash",
""
],
[
"György",
"András",
""
],
[
"Szepesvári",
"Csaba",
""
],
[
"Xu",
"Wilsun",
""
]
] | TITLE: SDP Relaxation with Randomized Rounding for Energy Disaggregation
ABSTRACT: We develop a scalable, computationally efficient method for the task of
energy disaggregation for home appliance monitoring. In this problem the goal
is to estimate the energy consumption of each appliance over time based on the
total energy-consumption signal of a household. The current state of the art is
to model the problem as inference in factorial HMMs, and use quadratic
programming to find an approximate solution to the resulting quadratic integer
program. Here we take a more principled approach, better suited to integer
programming problems, and find an approximate optimum by combining convex
semidefinite relaxations randomized rounding, as well as a scalable ADMM method
that exploits the special structure of the resulting semidefinite program.
Simulation results both in synthetic and real-world datasets demonstrate the
superiority of our method.
| no_new_dataset | 0.941601 |
1610.09500 | Yiming Lin | Yiming Lin and Hongzhi Wang and Jianzhong Li and Hong Gao | Efficient Entity Resolution on Heterogeneous Records | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Entity resolution (ER) is the problem of identifying and merging records that
refer to the same real-world entity. In many scenarios, raw records are stored
under heterogeneous environment. Specifically, the schemas of records may
differ from each other. To leverage such records better, most existing work
assume that schema matching and data exchange have been done to convert records
under different schemas to those under a predefined schema. However, we observe
that schema matching would lose information in some cases, which could be
useful or even crucial to ER.
To leverage sufficient information from heterogeneous sources, in this paper,
we address several challenges of ER on heterogeneous records and show that none
of existing similarity metrics or their transformations could be applied to
find similar records under heterogeneous settings. Motivated by this, we design
the similarity function and propose a novel framework to iteratively find
records which refer to the same entity. Regarding efficiency, we build an index
to generate candidates and accelerate similarity computation. Evaluations on
real-world datasets show the effectiveness and efficiency of our methods.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 12:51:52 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Lin",
"Yiming",
""
],
[
"Wang",
"Hongzhi",
""
],
[
"Li",
"Jianzhong",
""
],
[
"Gao",
"Hong",
""
]
] | TITLE: Efficient Entity Resolution on Heterogeneous Records
ABSTRACT: Entity resolution (ER) is the problem of identifying and merging records that
refer to the same real-world entity. In many scenarios, raw records are stored
under heterogeneous environment. Specifically, the schemas of records may
differ from each other. To leverage such records better, most existing work
assume that schema matching and data exchange have been done to convert records
under different schemas to those under a predefined schema. However, we observe
that schema matching would lose information in some cases, which could be
useful or even crucial to ER.
To leverage sufficient information from heterogeneous sources, in this paper,
we address several challenges of ER on heterogeneous records and show that none
of existing similarity metrics or their transformations could be applied to
find similar records under heterogeneous settings. Motivated by this, we design
the similarity function and propose a novel framework to iteratively find
records which refer to the same entity. Regarding efficiency, we build an index
to generate candidates and accelerate similarity computation. Evaluations on
real-world datasets show the effectiveness and efficiency of our methods.
| no_new_dataset | 0.950319 |
1610.09506 | Yiming Lin | Yiming Lin and Hongzhi Wang and Jianzhong Li and Hong Gao | Data Source Selection for Information Integration in Big Data Era | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In Big data era, information integration often requires abundant data
extracted from massive data sources. Due to a large number of data sources,
data source selection plays a crucial role in information integration, since it
is costly and even impossible to access all data sources. Data Source selection
should consider both efficiency and effectiveness issues. For efficiency, the
approach should achieve high performance and be scalability to fit large data
source amount. From effectiveness aspect, data quality and overlapping of
sources are to be considered, since data quality varies much from data sources,
with significant differences in the accuracy and coverage of the data provided,
and the overlapping of sources can even lower the quality of data integrated
from selected data sources.
In this paper, we study source selection problem in \textit{Big Data Era} and
propose methods which can scale to datasets with up to millions of data sources
and guarantee the quality of results. Motivated by this, we propose a new
object function taking the expected number of true values a source can provide
as a criteria to evaluate the contribution of a data source. Based on our
proposed index we present a scalable algorithm and two pruning strategies to
improve the efficiency without sacrificing precision. Experimental results on
both real world and synthetic data sets show that our methods can select
sources providing a large proportion of true values efficiently and can scale
to massive data sources.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 13:17:50 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Lin",
"Yiming",
""
],
[
"Wang",
"Hongzhi",
""
],
[
"Li",
"Jianzhong",
""
],
[
"Gao",
"Hong",
""
]
] | TITLE: Data Source Selection for Information Integration in Big Data Era
ABSTRACT: In Big data era, information integration often requires abundant data
extracted from massive data sources. Due to a large number of data sources,
data source selection plays a crucial role in information integration, since it
is costly and even impossible to access all data sources. Data Source selection
should consider both efficiency and effectiveness issues. For efficiency, the
approach should achieve high performance and be scalability to fit large data
source amount. From effectiveness aspect, data quality and overlapping of
sources are to be considered, since data quality varies much from data sources,
with significant differences in the accuracy and coverage of the data provided,
and the overlapping of sources can even lower the quality of data integrated
from selected data sources.
In this paper, we study source selection problem in \textit{Big Data Era} and
propose methods which can scale to datasets with up to millions of data sources
and guarantee the quality of results. Motivated by this, we propose a new
object function taking the expected number of true values a source can provide
as a criteria to evaluate the contribution of a data source. Based on our
proposed index we present a scalable algorithm and two pruning strategies to
improve the efficiency without sacrificing precision. Experimental results on
both real world and synthetic data sets show that our methods can select
sources providing a large proportion of true values efficiently and can scale
to massive data sources.
| no_new_dataset | 0.95018 |
1610.09565 | Mihaela Rosca | Mihaela Rosca, Thomas Breuel | Sequence-to-sequence neural network models for transliteration | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transliteration is a key component of machine translation systems and
software internationalization. This paper demonstrates that neural
sequence-to-sequence models obtain state of the art or close to state of the
art results on existing datasets. In an effort to make machine transliteration
accessible, we open source a new Arabic to English transliteration dataset and
our trained models.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 19:21:19 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Rosca",
"Mihaela",
""
],
[
"Breuel",
"Thomas",
""
]
] | TITLE: Sequence-to-sequence neural network models for transliteration
ABSTRACT: Transliteration is a key component of machine translation systems and
software internationalization. This paper demonstrates that neural
sequence-to-sequence models obtain state of the art or close to state of the
art results on existing datasets. In an effort to make machine transliteration
accessible, we open source a new Arabic to English transliteration dataset and
our trained models.
| new_dataset | 0.954308 |
1610.09582 | Rushil Anirudh | Rushil Anirudh, Ahnaf Masroor, Pavan Turaga | Diversity Promoting Online Sampling for Streaming Video Summarization | Published at ICIP 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many applications benefit from sampling algorithms where a small number of
well chosen samples are used to generalize different properties of a large
dataset. In this paper, we use diverse sampling for streaming video
summarization. Several emerging applications support streaming video, but
existing summarization algorithms need access to the entire video which
requires a lot of memory and computational power. We propose a memory efficient
and computationally fast, online algorithm that uses competitive learning for
diverse sampling. Our algorithm is a generalization of online K-means such that
the cost function reduces clustering error, while also ensuring a diverse set
of samples. The diversity is measured as the volume of a convex hull around the
samples. Finally, the performance of the proposed algorithm is measured against
human users for 50 videos in the VSUMM dataset. The algorithm performs better
than batch mode summarization, while requiring significantly lower memory and
computational requirements.
| [
{
"version": "v1",
"created": "Sat, 29 Oct 2016 23:51:24 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Anirudh",
"Rushil",
""
],
[
"Masroor",
"Ahnaf",
""
],
[
"Turaga",
"Pavan",
""
]
] | TITLE: Diversity Promoting Online Sampling for Streaming Video Summarization
ABSTRACT: Many applications benefit from sampling algorithms where a small number of
well chosen samples are used to generalize different properties of a large
dataset. In this paper, we use diverse sampling for streaming video
summarization. Several emerging applications support streaming video, but
existing summarization algorithms need access to the entire video which
requires a lot of memory and computational power. We propose a memory efficient
and computationally fast, online algorithm that uses competitive learning for
diverse sampling. Our algorithm is a generalization of online K-means such that
the cost function reduces clustering error, while also ensuring a diverse set
of samples. The diversity is measured as the volume of a convex hull around the
samples. Finally, the performance of the proposed algorithm is measured against
human users for 50 videos in the VSUMM dataset. The algorithm performs better
than batch mode summarization, while requiring significantly lower memory and
computational requirements.
| no_new_dataset | 0.947332 |
1610.09615 | Amir Adler | Amir Adler, Michael Elad and Michael Zibulevsky | Compressed Learning: A Deep Neural Network Approach | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Compressed Learning (CL) is a joint signal processing and machine learning
framework for inference from a signal, using a small number of measurements
obtained by linear projections of the signal. In this paper we present an
end-to-end deep learning approach for CL, in which a network composed of
fully-connected layers followed by convolutional layers perform the linear
sensing and non-linear inference stages. During the training phase, the sensing
matrix and the non-linear inference operator are jointly optimized, and the
proposed approach outperforms state-of-the-art for the task of image
classification. For example, at a sensing rate of 1% (only 8 measurements of 28
X 28 pixels images), the classification error for the MNIST handwritten digits
dataset is 6.46% compared to 41.06% with state-of-the-art.
| [
{
"version": "v1",
"created": "Sun, 30 Oct 2016 07:54:19 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Adler",
"Amir",
""
],
[
"Elad",
"Michael",
""
],
[
"Zibulevsky",
"Michael",
""
]
] | TITLE: Compressed Learning: A Deep Neural Network Approach
ABSTRACT: Compressed Learning (CL) is a joint signal processing and machine learning
framework for inference from a signal, using a small number of measurements
obtained by linear projections of the signal. In this paper we present an
end-to-end deep learning approach for CL, in which a network composed of
fully-connected layers followed by convolutional layers perform the linear
sensing and non-linear inference stages. During the training phase, the sensing
matrix and the non-linear inference operator are jointly optimized, and the
proposed approach outperforms state-of-the-art for the task of image
classification. For example, at a sensing rate of 1% (only 8 measurements of 28
X 28 pixels images), the classification error for the MNIST handwritten digits
dataset is 6.46% compared to 41.06% with state-of-the-art.
| no_new_dataset | 0.947186 |
1610.09639 | Sajid Anwar | Sajid Anwar, Wonyong Sung | Compact Deep Convolutional Neural Networks With Coarse Pruning | null | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The learning capability of a neural network improves with increasing depth at
higher computational costs. Wider layers with dense kernel connectivity
patterns furhter increase this cost and may hinder real-time inference. We
propose feature map and kernel level pruning for reducing the computational
complexity of a deep convolutional neural network. Pruning feature maps reduces
the width of a layer and hence does not need any sparse representation.
Further, kernel pruning converts the dense connectivity pattern into a sparse
one. Due to coarse nature, these pruning granularities can be exploited by GPUs
and VLSI based implementations. We propose a simple and generic strategy to
choose the least adversarial pruning masks for both granularities. The pruned
networks are retrained which compensates the loss in accuracy. We obtain the
best pruning ratios when we prune a network with both granularities.
Experiments with the CIFAR-10 dataset show that more than 85% sparsity can be
induced in the convolution layers with less than 1% increase in the
missclassification rate of the baseline network.
| [
{
"version": "v1",
"created": "Sun, 30 Oct 2016 11:57:20 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Anwar",
"Sajid",
""
],
[
"Sung",
"Wonyong",
""
]
] | TITLE: Compact Deep Convolutional Neural Networks With Coarse Pruning
ABSTRACT: The learning capability of a neural network improves with increasing depth at
higher computational costs. Wider layers with dense kernel connectivity
patterns furhter increase this cost and may hinder real-time inference. We
propose feature map and kernel level pruning for reducing the computational
complexity of a deep convolutional neural network. Pruning feature maps reduces
the width of a layer and hence does not need any sparse representation.
Further, kernel pruning converts the dense connectivity pattern into a sparse
one. Due to coarse nature, these pruning granularities can be exploited by GPUs
and VLSI based implementations. We propose a simple and generic strategy to
choose the least adversarial pruning masks for both granularities. The pruned
networks are retrained which compensates the loss in accuracy. We obtain the
best pruning ratios when we prune a network with both granularities.
Experiments with the CIFAR-10 dataset show that more than 85% sparsity can be
induced in the convolution layers with less than 1% increase in the
missclassification rate of the baseline network.
| no_new_dataset | 0.950915 |
1610.09652 | Kaihua Zhang | Kaihua Zhang, Qingshan Liu, and Ming-Hsuan Yang | Visual Tracking via Boolean Map Representations | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a simple yet effective Boolean map based
representation that exploits connectivity cues for visual tracking. We describe
a target object with histogram of oriented gradients and raw color features, of
which each one is characterized by a set of Boolean maps generated by uniformly
thresholding their values. The Boolean maps effectively encode multi-scale
connectivity cues of the target with different granularities. The fine-grained
Boolean maps capture spatially structural details that are effective for
precise target localization while the coarse-grained ones encode global shape
information that are robust to large target appearance variations. Finally, all
the Boolean maps form together a robust representation that can be approximated
by an explicit feature map of the intersection kernel, which is fed into a
logistic regression classifier with online update, and the target location is
estimated within a particle filter framework. The proposed representation
scheme is computationally efficient and facilitates achieving favorable
performance in terms of accuracy and robustness against the state-of-the-art
tracking methods on a large benchmark dataset of 50 image sequences.
| [
{
"version": "v1",
"created": "Sun, 30 Oct 2016 14:17:05 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Zhang",
"Kaihua",
""
],
[
"Liu",
"Qingshan",
""
],
[
"Yang",
"Ming-Hsuan",
""
]
] | TITLE: Visual Tracking via Boolean Map Representations
ABSTRACT: In this paper, we present a simple yet effective Boolean map based
representation that exploits connectivity cues for visual tracking. We describe
a target object with histogram of oriented gradients and raw color features, of
which each one is characterized by a set of Boolean maps generated by uniformly
thresholding their values. The Boolean maps effectively encode multi-scale
connectivity cues of the target with different granularities. The fine-grained
Boolean maps capture spatially structural details that are effective for
precise target localization while the coarse-grained ones encode global shape
information that are robust to large target appearance variations. Finally, all
the Boolean maps form together a robust representation that can be approximated
by an explicit feature map of the intersection kernel, which is fed into a
logistic regression classifier with online update, and the target location is
estimated within a particle filter framework. The proposed representation
scheme is computationally efficient and facilitates achieving favorable
performance in terms of accuracy and robustness against the state-of-the-art
tracking methods on a large benchmark dataset of 50 image sequences.
| no_new_dataset | 0.949059 |
1610.09778 | Jingbo Shang | Jingbo Shang, Meng Jiang, Wenzhu Tong, Jinfeng Xiao, Jian Peng, Jiawei
Han | DPPred: An Effective Prediction Framework with Concise Discriminative
Patterns | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the literature, two series of models have been proposed to address
prediction problems including classification and regression. Simple models,
such as generalized linear models, have ordinary performance but strong
interpretability on a set of simple features. The other series, including
tree-based models, organize numerical, categorical and high dimensional
features into a comprehensive structure with rich interpretable information in
the data.
In this paper, we propose a novel Discriminative Pattern-based Prediction
framework (DPPred) to accomplish the prediction tasks by taking their
advantages of both effectiveness and interpretability. Specifically, DPPred
adopts the concise discriminative patterns that are on the prefix paths from
the root to leaf nodes in the tree-based models. DPPred selects a limited
number of the useful discriminative patterns by searching for the most
effective pattern combination to fit generalized linear models. Extensive
experiments show that in many scenarios, DPPred provides competitive accuracy
with the state-of-the-art as well as the valuable interpretability for
developers and experts. In particular, taking a clinical application dataset as
a case study, our DPPred outperforms the baselines by using only 40 concise
discriminative patterns out of a potentially exponentially large set of
patterns.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 03:43:04 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Shang",
"Jingbo",
""
],
[
"Jiang",
"Meng",
""
],
[
"Tong",
"Wenzhu",
""
],
[
"Xiao",
"Jinfeng",
""
],
[
"Peng",
"Jian",
""
],
[
"Han",
"Jiawei",
""
]
] | TITLE: DPPred: An Effective Prediction Framework with Concise Discriminative
Patterns
ABSTRACT: In the literature, two series of models have been proposed to address
prediction problems including classification and regression. Simple models,
such as generalized linear models, have ordinary performance but strong
interpretability on a set of simple features. The other series, including
tree-based models, organize numerical, categorical and high dimensional
features into a comprehensive structure with rich interpretable information in
the data.
In this paper, we propose a novel Discriminative Pattern-based Prediction
framework (DPPred) to accomplish the prediction tasks by taking their
advantages of both effectiveness and interpretability. Specifically, DPPred
adopts the concise discriminative patterns that are on the prefix paths from
the root to leaf nodes in the tree-based models. DPPred selects a limited
number of the useful discriminative patterns by searching for the most
effective pattern combination to fit generalized linear models. Extensive
experiments show that in many scenarios, DPPred provides competitive accuracy
with the state-of-the-art as well as the valuable interpretability for
developers and experts. In particular, taking a clinical application dataset as
a case study, our DPPred outperforms the baselines by using only 40 concise
discriminative patterns out of a potentially exponentially large set of
patterns.
| no_new_dataset | 0.944893 |
1610.09893 | Tao Qin Dr. | Xiang Li and Tao Qin and Jian Yang and Tie-Yan Liu | LightRNN: Memory and Computation-Efficient Recurrent Neural Networks | NIPS 2016 | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recurrent neural networks (RNNs) have achieved state-of-the-art performances
in many natural language processing tasks, such as language modeling and
machine translation. However, when the vocabulary is large, the RNN model will
become very big (e.g., possibly beyond the memory capacity of a GPU device) and
its training will become very inefficient. In this work, we propose a novel
technique to tackle this challenge. The key idea is to use 2-Component (2C)
shared embedding for word representations. We allocate every word in the
vocabulary into a table, each row of which is associated with a vector, and
each column associated with another vector. Depending on its position in the
table, a word is jointly represented by two components: a row vector and a
column vector. Since the words in the same row share the row vector and the
words in the same column share the column vector, we only need $2 \sqrt{|V|}$
vectors to represent a vocabulary of $|V|$ unique words, which are far less
than the $|V|$ vectors required by existing approaches. Based on the
2-Component shared embedding, we design a new RNN algorithm and evaluate it
using the language modeling task on several benchmark datasets. The results
show that our algorithm significantly reduces the model size and speeds up the
training process, without sacrifice of accuracy (it achieves similar, if not
better, perplexity as compared to state-of-the-art language models).
Remarkably, on the One-Billion-Word benchmark Dataset, our algorithm achieves
comparable perplexity to previous language models, whilst reducing the model
size by a factor of 40-100, and speeding up the training process by a factor of
2. We name our proposed algorithm \emph{LightRNN} to reflect its very small
model size and very high training speed.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 12:24:13 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Li",
"Xiang",
""
],
[
"Qin",
"Tao",
""
],
[
"Yang",
"Jian",
""
],
[
"Liu",
"Tie-Yan",
""
]
] | TITLE: LightRNN: Memory and Computation-Efficient Recurrent Neural Networks
ABSTRACT: Recurrent neural networks (RNNs) have achieved state-of-the-art performances
in many natural language processing tasks, such as language modeling and
machine translation. However, when the vocabulary is large, the RNN model will
become very big (e.g., possibly beyond the memory capacity of a GPU device) and
its training will become very inefficient. In this work, we propose a novel
technique to tackle this challenge. The key idea is to use 2-Component (2C)
shared embedding for word representations. We allocate every word in the
vocabulary into a table, each row of which is associated with a vector, and
each column associated with another vector. Depending on its position in the
table, a word is jointly represented by two components: a row vector and a
column vector. Since the words in the same row share the row vector and the
words in the same column share the column vector, we only need $2 \sqrt{|V|}$
vectors to represent a vocabulary of $|V|$ unique words, which are far less
than the $|V|$ vectors required by existing approaches. Based on the
2-Component shared embedding, we design a new RNN algorithm and evaluate it
using the language modeling task on several benchmark datasets. The results
show that our algorithm significantly reduces the model size and speeds up the
training process, without sacrifice of accuracy (it achieves similar, if not
better, perplexity as compared to state-of-the-art language models).
Remarkably, on the One-Billion-Word benchmark Dataset, our algorithm achieves
comparable perplexity to previous language models, whilst reducing the model
size by a factor of 40-100, and speeding up the training process by a factor of
2. We name our proposed algorithm \emph{LightRNN} to reflect its very small
model size and very high training speed.
| no_new_dataset | 0.948728 |
1610.09984 | Alessandro Epasto | Alessandro Epasto, Silvio Lattanzi, Sergei Vassilvitskii, Morteza
Zadimoghaddam | Submodular Optimization over Sliding Windows | null | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Maximizing submodular functions under cardinality constraints lies at the
core of numerous data mining and machine learning applications, including data
diversification, data summarization, and coverage problems. In this work, we
study this question in the context of data streams, where elements arrive one
at a time, and we want to design low-memory and fast update-time algorithms
that maintain a good solution. Specifically, we focus on the sliding window
model, where we are asked to maintain a solution that considers only the last
$W$ items.
In this context, we provide the first non-trivial algorithm that maintains a
provable approximation of the optimum using space sublinear in the size of the
window. In particular we give a $\frac{1}{3} - \epsilon$ approximation
algorithm that uses space polylogarithmic in the spread of the values of the
elements, $\Phi$, and linear in the solution size $k$ for any constant
$\epsilon > 0$ . At the same time, processing each element only requires a
polylogarithmic number of evaluations of the function itself. When a better
approximation is desired, we show a different algorithm that, at the cost of
using more memory, provides a $\frac{1}{2} - \epsilon$ approximation and allows
a tunable trade-off between average update time and space. This algorithm
matches the best known approximation guarantees for submodular optimization in
insertion-only streams, a less general formulation of the problem.
We demonstrate the efficacy of the algorithms on a number of real world
datasets, showing that their practical performance far exceeds the theoretical
bounds. The algorithms preserve high quality solutions in streams with millions
of items, while storing a negligible fraction of them.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 15:48:24 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Epasto",
"Alessandro",
""
],
[
"Lattanzi",
"Silvio",
""
],
[
"Vassilvitskii",
"Sergei",
""
],
[
"Zadimoghaddam",
"Morteza",
""
]
] | TITLE: Submodular Optimization over Sliding Windows
ABSTRACT: Maximizing submodular functions under cardinality constraints lies at the
core of numerous data mining and machine learning applications, including data
diversification, data summarization, and coverage problems. In this work, we
study this question in the context of data streams, where elements arrive one
at a time, and we want to design low-memory and fast update-time algorithms
that maintain a good solution. Specifically, we focus on the sliding window
model, where we are asked to maintain a solution that considers only the last
$W$ items.
In this context, we provide the first non-trivial algorithm that maintains a
provable approximation of the optimum using space sublinear in the size of the
window. In particular we give a $\frac{1}{3} - \epsilon$ approximation
algorithm that uses space polylogarithmic in the spread of the values of the
elements, $\Phi$, and linear in the solution size $k$ for any constant
$\epsilon > 0$ . At the same time, processing each element only requires a
polylogarithmic number of evaluations of the function itself. When a better
approximation is desired, we show a different algorithm that, at the cost of
using more memory, provides a $\frac{1}{2} - \epsilon$ approximation and allows
a tunable trade-off between average update time and space. This algorithm
matches the best known approximation guarantees for submodular optimization in
insertion-only streams, a less general formulation of the problem.
We demonstrate the efficacy of the algorithms on a number of real world
datasets, showing that their practical performance far exceeds the theoretical
bounds. The algorithms preserve high quality solutions in streams with millions
of items, while storing a negligible fraction of them.
| no_new_dataset | 0.940298 |
1610.10048 | Arulkumar Subramaniam | Arulkumar Subramaniam, Vismay Patel, Ashish Mishra, Prashanth
Balasubramanian, Anurag Mittal | Bi-modal First Impressions Recognition using Temporally Ordered Deep
Audio and Stochastic Visual Features | to be published in: ECCV 2016 Workshops proceedings (Apparent
Personality Analysis) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel approach for First Impressions Recognition in terms of the
Big Five personality-traits from short videos. The Big Five personality traits
is a model to describe human personality using five broad categories:
Extraversion, Agreeableness, Conscientiousness, Neuroticism and Openness. We
train two bi-modal end-to-end deep neural network architectures using
temporally ordered audio and novel stochastic visual features from few frames,
without over-fitting. We empirically show that the trained models perform
exceptionally well, even after training from a small sub-portions of inputs.
Our method is evaluated in ChaLearn LAP 2016 Apparent Personality Analysis
(APA) competition using ChaLearn LAP APA2016 dataset and achieved excellent
performance.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 18:21:13 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Subramaniam",
"Arulkumar",
""
],
[
"Patel",
"Vismay",
""
],
[
"Mishra",
"Ashish",
""
],
[
"Balasubramanian",
"Prashanth",
""
],
[
"Mittal",
"Anurag",
""
]
] | TITLE: Bi-modal First Impressions Recognition using Temporally Ordered Deep
Audio and Stochastic Visual Features
ABSTRACT: We propose a novel approach for First Impressions Recognition in terms of the
Big Five personality-traits from short videos. The Big Five personality traits
is a model to describe human personality using five broad categories:
Extraversion, Agreeableness, Conscientiousness, Neuroticism and Openness. We
train two bi-modal end-to-end deep neural network architectures using
temporally ordered audio and novel stochastic visual features from few frames,
without over-fitting. We empirically show that the trained models perform
exceptionally well, even after training from a small sub-portions of inputs.
Our method is evaluated in ChaLearn LAP 2016 Apparent Personality Analysis
(APA) competition using ChaLearn LAP APA2016 dataset and achieved excellent
performance.
| no_new_dataset | 0.951233 |
1610.10064 | Muhammad Bilal Zafar | Miguel Ferreira, Muhammad Bilal Zafar, Krishna P. Gummadi | The Case for Temporal Transparency: Detecting Policy Change Events in
Black-Box Decision Making Systems | null | null | null | null | stat.ML cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bringing transparency to black-box decision making systems (DMS) has been a
topic of increasing research interest in recent years. Traditional active and
passive approaches to make these systems transparent are often limited by
scalability and/or feasibility issues. In this paper, we propose a new notion
of black-box DMS transparency, named, temporal transparency, whose goal is to
detect if/when the DMS policy changes over time, and is mostly invariant to the
drawbacks of traditional approaches. We map our notion of temporal transparency
to time series changepoint detection methods, and develop a framework to detect
policy changes in real-world DMS's. Experiments on New York
Stop-question-and-frisk dataset reveal a number of publicly announced and
unannounced policy changes, highlighting the utility of our framework.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2016 18:54:56 GMT"
}
] | 2016-11-01T00:00:00 | [
[
"Ferreira",
"Miguel",
""
],
[
"Zafar",
"Muhammad Bilal",
""
],
[
"Gummadi",
"Krishna P.",
""
]
] | TITLE: The Case for Temporal Transparency: Detecting Policy Change Events in
Black-Box Decision Making Systems
ABSTRACT: Bringing transparency to black-box decision making systems (DMS) has been a
topic of increasing research interest in recent years. Traditional active and
passive approaches to make these systems transparent are often limited by
scalability and/or feasibility issues. In this paper, we propose a new notion
of black-box DMS transparency, named, temporal transparency, whose goal is to
detect if/when the DMS policy changes over time, and is mostly invariant to the
drawbacks of traditional approaches. We map our notion of temporal transparency
to time series changepoint detection methods, and develop a framework to detect
policy changes in real-world DMS's. Experiments on New York
Stop-question-and-frisk dataset reveal a number of publicly announced and
unannounced policy changes, highlighting the utility of our framework.
| no_new_dataset | 0.954393 |
1409.5253 | Valerio Gemmetto | Valerio Gemmetto and Diego Garlaschelli | Multiplexity versus correlation: the role of local constraints in real
multiplexes | 32 pages, 6 figures | Scientific Reports 5, 9120 (2015) | 10.1038/srep09120 | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several real-world systems can be represented as multi-layer complex
networks, i.e. in terms of a superposition of various graphs, each related to a
different mode of connection between nodes. Hence, the definition of proper
mathematical quantities aiming at capturing the level of complexity of those
systems is required. Various attempts have been made to measure the empirical
dependencies between the layers of a multiplex, for both binary and weighted
networks. In the simplest case, such dependencies are measured via
correlation-based metrics: we show that this is equivalent to the use of
completely homogeneous benchmarks specifying only global constraints, such as
the total number of links in each layer. However, these approaches do not take
into account the heterogeneity in the degree and strength distributions, which
are instead a fundamental feature of real-world multiplexes. In this work, we
compare the observed dependencies between layers with the expected values
obtained from reference models that appropriately control for the observed
heterogeneity in the degree and strength distributions. This leads to novel
multiplexity measures that we test on different datasets, i.e. the
International Trade Network (ITN) and the European Airport Network (EAN). Our
findings confirm that the use of homogeneous benchmarks can lead to misleading
results, and furthermore highlight the important role played by the
distribution of hubs across layers.
| [
{
"version": "v1",
"created": "Thu, 18 Sep 2014 10:49:45 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Gemmetto",
"Valerio",
""
],
[
"Garlaschelli",
"Diego",
""
]
] | TITLE: Multiplexity versus correlation: the role of local constraints in real
multiplexes
ABSTRACT: Several real-world systems can be represented as multi-layer complex
networks, i.e. in terms of a superposition of various graphs, each related to a
different mode of connection between nodes. Hence, the definition of proper
mathematical quantities aiming at capturing the level of complexity of those
systems is required. Various attempts have been made to measure the empirical
dependencies between the layers of a multiplex, for both binary and weighted
networks. In the simplest case, such dependencies are measured via
correlation-based metrics: we show that this is equivalent to the use of
completely homogeneous benchmarks specifying only global constraints, such as
the total number of links in each layer. However, these approaches do not take
into account the heterogeneity in the degree and strength distributions, which
are instead a fundamental feature of real-world multiplexes. In this work, we
compare the observed dependencies between layers with the expected values
obtained from reference models that appropriately control for the observed
heterogeneity in the degree and strength distributions. This leads to novel
multiplexity measures that we test on different datasets, i.e. the
International Trade Network (ITN) and the European Airport Network (EAN). Our
findings confirm that the use of homogeneous benchmarks can lead to misleading
results, and furthermore highlight the important role played by the
distribution of hubs across layers.
| no_new_dataset | 0.944893 |
1601.06180 | Robert Peharz | Robert Peharz, Robert Gens, Franz Pernkopf, Pedro Domingos | On the Latent Variable Interpretation in Sum-Product Networks | Revised version, accepted for publication in IEEE Transactions on
Machine Intelligence and Pattern Analysis (TPAMI). Shortened and revised
Section 4: Thanks to our reviewers, pointing out that Theorem 2 holds for
selective SPNs. Added paragraph in Section 2.1, relating sizes of
original/augmented SPNs. Fixed typos, rephrased sentences, revised references | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the central themes in Sum-Product networks (SPNs) is the
interpretation of sum nodes as marginalized latent variables (LVs). This
interpretation yields an increased syntactic or semantic structure, allows the
application of the EM algorithm and to efficiently perform MPE inference. In
literature, the LV interpretation was justified by explicitly introducing the
indicator variables corresponding to the LVs' states. However, as pointed out
in this paper, this approach is in conflict with the completeness condition in
SPNs and does not fully specify the probabilistic model. We propose a remedy
for this problem by modifying the original approach for introducing the LVs,
which we call SPN augmentation. We discuss conditional independencies in
augmented SPNs, formally establish the probabilistic interpretation of the
sum-weights and give an interpretation of augmented SPNs as Bayesian networks.
Based on these results, we find a sound derivation of the EM algorithm for
SPNs. Furthermore, the Viterbi-style algorithm for MPE proposed in literature
was never proven to be correct. We show that this is indeed a correct
algorithm, when applied to selective SPNs, and in particular when applied to
augmented SPNs. Our theoretical results are confirmed in experiments on
synthetic data and 103 real-world datasets.
| [
{
"version": "v1",
"created": "Fri, 22 Jan 2016 21:40:33 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Oct 2016 07:54:35 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Peharz",
"Robert",
""
],
[
"Gens",
"Robert",
""
],
[
"Pernkopf",
"Franz",
""
],
[
"Domingos",
"Pedro",
""
]
] | TITLE: On the Latent Variable Interpretation in Sum-Product Networks
ABSTRACT: One of the central themes in Sum-Product networks (SPNs) is the
interpretation of sum nodes as marginalized latent variables (LVs). This
interpretation yields an increased syntactic or semantic structure, allows the
application of the EM algorithm and to efficiently perform MPE inference. In
literature, the LV interpretation was justified by explicitly introducing the
indicator variables corresponding to the LVs' states. However, as pointed out
in this paper, this approach is in conflict with the completeness condition in
SPNs and does not fully specify the probabilistic model. We propose a remedy
for this problem by modifying the original approach for introducing the LVs,
which we call SPN augmentation. We discuss conditional independencies in
augmented SPNs, formally establish the probabilistic interpretation of the
sum-weights and give an interpretation of augmented SPNs as Bayesian networks.
Based on these results, we find a sound derivation of the EM algorithm for
SPNs. Furthermore, the Viterbi-style algorithm for MPE proposed in literature
was never proven to be correct. We show that this is indeed a correct
algorithm, when applied to selective SPNs, and in particular when applied to
augmented SPNs. Our theoretical results are confirmed in experiments on
synthetic data and 103 real-world datasets.
| no_new_dataset | 0.949995 |
1606.08513 | Tomasz Jurczyk | Tomasz Jurczyk, Michael Zhai, Jinho D. Choi | SelQA: A New Benchmark for Selection-based Question Answering | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a new selection-based question answering dataset, SelQA.
The dataset consists of questions generated through crowdsourcing and sentence
length answers that are drawn from the ten most prevalent topics in the English
Wikipedia. We introduce a corpus annotation scheme that enhances the generation
of large, diverse, and challenging datasets by explicitly aiming to reduce word
co-occurrences between the question and answers. Our annotation scheme is
composed of a series of crowdsourcing tasks with a view to more effectively
utilize crowdsourcing in the creation of question answering datasets in various
domains. Several systems are compared on the tasks of answer sentence selection
and answer triggering, providing strong baseline results for future work to
improve upon.
| [
{
"version": "v1",
"created": "Mon, 27 Jun 2016 23:48:16 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Oct 2016 16:36:02 GMT"
},
{
"version": "v3",
"created": "Fri, 28 Oct 2016 01:20:19 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Jurczyk",
"Tomasz",
""
],
[
"Zhai",
"Michael",
""
],
[
"Choi",
"Jinho D.",
""
]
] | TITLE: SelQA: A New Benchmark for Selection-based Question Answering
ABSTRACT: This paper presents a new selection-based question answering dataset, SelQA.
The dataset consists of questions generated through crowdsourcing and sentence
length answers that are drawn from the ten most prevalent topics in the English
Wikipedia. We introduce a corpus annotation scheme that enhances the generation
of large, diverse, and challenging datasets by explicitly aiming to reduce word
co-occurrences between the question and answers. Our annotation scheme is
composed of a series of crowdsourcing tasks with a view to more effectively
utilize crowdsourcing in the creation of question answering datasets in various
domains. Several systems are compared on the tasks of answer sentence selection
and answer triggering, providing strong baseline results for future work to
improve upon.
| new_dataset | 0.952706 |
1607.02046 | Gr\'egory Rogez | Gr\'egory Rogez and Cordelia Schmid | MoCap-guided Data Augmentation for 3D Pose Estimation in the Wild | 9 pages, accepted to appear in NIPS 2016 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of 3D human pose estimation in the wild. A
significant challenge is the lack of training data, i.e., 2D images of humans
annotated with 3D poses. Such data is necessary to train state-of-the-art CNN
architectures. Here, we propose a solution to generate a large set of
photorealistic synthetic images of humans with 3D pose annotations. We
introduce an image-based synthesis engine that artificially augments a dataset
of real images with 2D human pose annotations using 3D Motion Capture (MoCap)
data. Given a candidate 3D pose our algorithm selects for each joint an image
whose 2D pose locally matches the projected 3D pose. The selected images are
then combined to generate a new synthetic image by stitching local image
patches in a kinematically constrained manner. The resulting images are used to
train an end-to-end CNN for full-body 3D pose estimation. We cluster the
training data into a large number of pose classes and tackle pose estimation as
a K-way classification problem. Such an approach is viable only with large
training sets such as ours. Our method outperforms the state of the art in
terms of 3D pose estimation in controlled environments (Human3.6M) and shows
promising results for in-the-wild images (LSP). This demonstrates that CNNs
trained on artificial images generalize well to real images.
| [
{
"version": "v1",
"created": "Thu, 7 Jul 2016 15:30:05 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Oct 2016 12:43:51 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Rogez",
"Grégory",
""
],
[
"Schmid",
"Cordelia",
""
]
] | TITLE: MoCap-guided Data Augmentation for 3D Pose Estimation in the Wild
ABSTRACT: This paper addresses the problem of 3D human pose estimation in the wild. A
significant challenge is the lack of training data, i.e., 2D images of humans
annotated with 3D poses. Such data is necessary to train state-of-the-art CNN
architectures. Here, we propose a solution to generate a large set of
photorealistic synthetic images of humans with 3D pose annotations. We
introduce an image-based synthesis engine that artificially augments a dataset
of real images with 2D human pose annotations using 3D Motion Capture (MoCap)
data. Given a candidate 3D pose our algorithm selects for each joint an image
whose 2D pose locally matches the projected 3D pose. The selected images are
then combined to generate a new synthetic image by stitching local image
patches in a kinematically constrained manner. The resulting images are used to
train an end-to-end CNN for full-body 3D pose estimation. We cluster the
training data into a large number of pose classes and tackle pose estimation as
a K-way classification problem. Such an approach is viable only with large
training sets such as ours. Our method outperforms the state of the art in
terms of 3D pose estimation in controlled environments (Human3.6M) and shows
promising results for in-the-wild images (LSP). This demonstrates that CNNs
trained on artificial images generalize well to real images.
| no_new_dataset | 0.946051 |
1609.08777 | Kazuya Kawakami | Kazuya Kawakami, Chris Dyer, Bryan R. Routledge, Noah A. Smith | Character Sequence Models for ColorfulWords | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a neural network architecture to predict a point in color space
from the sequence of characters in the color's name. Using large scale
color--name pairs obtained from an online color design forum, we evaluate our
model on a "color Turing test" and find that, given a name, the colors
predicted by our model are preferred by annotators to color names created by
humans. Our datasets and demo system are available online at colorlab.us.
| [
{
"version": "v1",
"created": "Wed, 28 Sep 2016 05:41:18 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Oct 2016 16:08:36 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Kawakami",
"Kazuya",
""
],
[
"Dyer",
"Chris",
""
],
[
"Routledge",
"Bryan R.",
""
],
[
"Smith",
"Noah A.",
""
]
] | TITLE: Character Sequence Models for ColorfulWords
ABSTRACT: We present a neural network architecture to predict a point in color space
from the sequence of characters in the color's name. Using large scale
color--name pairs obtained from an online color design forum, we evaluate our
model on a "color Turing test" and find that, given a name, the colors
predicted by our model are preferred by annotators to color names created by
humans. Our datasets and demo system are available online at colorlab.us.
| no_new_dataset | 0.943295 |
1610.09003 | Carl Vondrick | Yusuf Aytar, Lluis Castrejon, Carl Vondrick, Hamed Pirsiavash, Antonio
Torralba | Cross-Modal Scene Networks | See more at http://cmplaces.csail.mit.edu/. arXiv admin note: text
overlap with arXiv:1607.07295 | null | null | null | cs.CV cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | People can recognize scenes across many different modalities beyond natural
images. In this paper, we investigate how to learn cross-modal scene
representations that transfer across modalities. To study this problem, we
introduce a new cross-modal scene dataset. While convolutional neural networks
can categorize scenes well, they also learn an intermediate representation not
aligned across modalities, which is undesirable for cross-modal transfer
applications. We present methods to regularize cross-modal convolutional neural
networks so that they have a shared representation that is agnostic of the
modality. Our experiments suggest that our scene representation can help
transfer representations across modalities for retrieval. Moreover, our
visualizations suggest that units emerge in the shared representation that tend
to activate on consistent concepts independently of the modality.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 20:24:36 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Aytar",
"Yusuf",
""
],
[
"Castrejon",
"Lluis",
""
],
[
"Vondrick",
"Carl",
""
],
[
"Pirsiavash",
"Hamed",
""
],
[
"Torralba",
"Antonio",
""
]
] | TITLE: Cross-Modal Scene Networks
ABSTRACT: People can recognize scenes across many different modalities beyond natural
images. In this paper, we investigate how to learn cross-modal scene
representations that transfer across modalities. To study this problem, we
introduce a new cross-modal scene dataset. While convolutional neural networks
can categorize scenes well, they also learn an intermediate representation not
aligned across modalities, which is undesirable for cross-modal transfer
applications. We present methods to regularize cross-modal convolutional neural
networks so that they have a shared representation that is agnostic of the
modality. Our experiments suggest that our scene representation can help
transfer representations across modalities for retrieval. Moreover, our
visualizations suggest that units emerge in the shared representation that tend
to activate on consistent concepts independently of the modality.
| new_dataset | 0.956796 |
1610.09058 | Riley Murray | Riley Murray and Samir Khuller and Megan Chao | Scheduling Distributed Clusters of Parallel Machines: Primal-Dual and
LP-based Approximation Algorithms [Full Version] | A shorter version of this paper (one that omitted several proofs)
appeared in the proceedings of the 2016 European Symposium on Algorithms | Leibniz International Proceedings in Informatics (LIPIcs), Volume
58, 2016, pages 68:1--68:17 | 10.4230/LIPIcs.ESA.2016.68 | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Map-Reduce computing framework rose to prominence with datasets of such
size that dozens of machines on a single cluster were needed for individual
jobs. As datasets approach the exabyte scale, a single job may need distributed
processing not only on multiple machines, but on multiple clusters. We consider
a scheduling problem to minimize weighted average completion time of N jobs on
M distributed clusters of parallel machines. In keeping with the scale of the
problems motivating this work, we assume that (1) each job is divided into M
"subjobs" and (2) distinct subjobs of a given job may be processed
concurrently.
When each cluster is a single machine, this is the NP-Hard concurrent open
shop problem. A clear limitation of such a model is that a serial processing
assumption sidesteps the issue of how different tasks of a given subjob might
be processed in parallel. Our algorithms explicitly model clusters as pools of
resources and effectively overcome this issue.
Under a variety of parameter settings, we develop two constant factor
approximation algorithms for this problem. The first algorithm uses an LP
relaxation tailored to this problem from prior work. This LP-based algorithm
provides strong performance guarantees. Our second algorithm exploits a
surprisingly simple mapping to the special case of one machine per cluster.
This mapping-based algorithm is combinatorial and extremely fast. These are the
first constant factor approximations for this problem.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2016 02:14:25 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Murray",
"Riley",
""
],
[
"Khuller",
"Samir",
""
],
[
"Chao",
"Megan",
""
]
] | TITLE: Scheduling Distributed Clusters of Parallel Machines: Primal-Dual and
LP-based Approximation Algorithms [Full Version]
ABSTRACT: The Map-Reduce computing framework rose to prominence with datasets of such
size that dozens of machines on a single cluster were needed for individual
jobs. As datasets approach the exabyte scale, a single job may need distributed
processing not only on multiple machines, but on multiple clusters. We consider
a scheduling problem to minimize weighted average completion time of N jobs on
M distributed clusters of parallel machines. In keeping with the scale of the
problems motivating this work, we assume that (1) each job is divided into M
"subjobs" and (2) distinct subjobs of a given job may be processed
concurrently.
When each cluster is a single machine, this is the NP-Hard concurrent open
shop problem. A clear limitation of such a model is that a serial processing
assumption sidesteps the issue of how different tasks of a given subjob might
be processed in parallel. Our algorithms explicitly model clusters as pools of
resources and effectively overcome this issue.
Under a variety of parameter settings, we develop two constant factor
approximation algorithms for this problem. The first algorithm uses an LP
relaxation tailored to this problem from prior work. This LP-based algorithm
provides strong performance guarantees. Our second algorithm exploits a
surprisingly simple mapping to the special case of one machine per cluster.
This mapping-based algorithm is combinatorial and extremely fast. These are the
first constant factor approximations for this problem.
| no_new_dataset | 0.942929 |
1610.09072 | Felix X. Yu | Felix X. Yu, Ananda Theertha Suresh, Krzysztof Choromanski, Daniel
Holtmann-Rice, Sanjiv Kumar | Orthogonal Random Features | NIPS 2016 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an intriguing discovery related to Random Fourier Features: in
Gaussian kernel approximation, replacing the random Gaussian matrix by a
properly scaled random orthogonal matrix significantly decreases kernel
approximation error. We call this technique Orthogonal Random Features (ORF),
and provide theoretical and empirical justification for this behavior.
Motivated by this discovery, we further propose Structured Orthogonal Random
Features (SORF), which uses a class of structured discrete orthogonal matrices
to speed up the computation. The method reduces the time cost from
$\mathcal{O}(d^2)$ to $\mathcal{O}(d \log d)$, where $d$ is the data
dimensionality, with almost no compromise in kernel approximation quality
compared to ORF. Experiments on several datasets verify the effectiveness of
ORF and SORF over the existing methods. We also provide discussions on using
the same type of discrete orthogonal structure for a broader range of
applications.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2016 03:50:00 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Yu",
"Felix X.",
""
],
[
"Suresh",
"Ananda Theertha",
""
],
[
"Choromanski",
"Krzysztof",
""
],
[
"Holtmann-Rice",
"Daniel",
""
],
[
"Kumar",
"Sanjiv",
""
]
] | TITLE: Orthogonal Random Features
ABSTRACT: We present an intriguing discovery related to Random Fourier Features: in
Gaussian kernel approximation, replacing the random Gaussian matrix by a
properly scaled random orthogonal matrix significantly decreases kernel
approximation error. We call this technique Orthogonal Random Features (ORF),
and provide theoretical and empirical justification for this behavior.
Motivated by this discovery, we further propose Structured Orthogonal Random
Features (SORF), which uses a class of structured discrete orthogonal matrices
to speed up the computation. The method reduces the time cost from
$\mathcal{O}(d^2)$ to $\mathcal{O}(d \log d)$, where $d$ is the data
dimensionality, with almost no compromise in kernel approximation quality
compared to ORF. Experiments on several datasets verify the effectiveness of
ORF and SORF over the existing methods. We also provide discussions on using
the same type of discrete orthogonal structure for a broader range of
applications.
| no_new_dataset | 0.953188 |
1610.09274 | Guang-He Lee | Guang-He Lee, Shao-Wen Yang, Shou-De Lin | Toward Implicit Sample Noise Modeling: Deviation-driven Matrix
Factorization | 6 pages + 1 reference page | null | null | null | cs.LG cs.IR stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The objective function of a matrix factorization model usually aims to
minimize the average of a regression error contributed by each element.
However, given the existence of stochastic noises, the implicit deviations of
sample data from their true values are almost surely diverse, which makes each
data point not equally suitable for fitting a model. In this case, simply
averaging the cost among data in the objective function is not ideal.
Intuitively we would like to emphasize more on the reliable instances (i.e.,
those contain smaller noise) while training a model. Motivated by such
observation, we derive our formula from a theoretical framework for optimal
weighting under heteroscedastic noise distribution. Specifically, by modeling
and learning the deviation of data, we design a novel matrix factorization
model. Our model has two advantages. First, it jointly learns the deviation and
conducts dynamic reweighting of instances, allowing the model to converge to a
better solution. Second, during learning the deviated instances are assigned
lower weights, which leads to faster convergence since the model does not need
to overfit the noise. The experiments are conducted in clean recommendation and
noisy sensor datasets to test the effectiveness of the model in various
scenarios. The results show that our model outperforms the state-of-the-art
factorization and deep learning models in both accuracy and efficiency.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2016 15:33:25 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Lee",
"Guang-He",
""
],
[
"Yang",
"Shao-Wen",
""
],
[
"Lin",
"Shou-De",
""
]
] | TITLE: Toward Implicit Sample Noise Modeling: Deviation-driven Matrix
Factorization
ABSTRACT: The objective function of a matrix factorization model usually aims to
minimize the average of a regression error contributed by each element.
However, given the existence of stochastic noises, the implicit deviations of
sample data from their true values are almost surely diverse, which makes each
data point not equally suitable for fitting a model. In this case, simply
averaging the cost among data in the objective function is not ideal.
Intuitively we would like to emphasize more on the reliable instances (i.e.,
those contain smaller noise) while training a model. Motivated by such
observation, we derive our formula from a theoretical framework for optimal
weighting under heteroscedastic noise distribution. Specifically, by modeling
and learning the deviation of data, we design a novel matrix factorization
model. Our model has two advantages. First, it jointly learns the deviation and
conducts dynamic reweighting of instances, allowing the model to converge to a
better solution. Second, during learning the deviated instances are assigned
lower weights, which leads to faster convergence since the model does not need
to overfit the noise. The experiments are conducted in clean recommendation and
noisy sensor datasets to test the effectiveness of the model in various
scenarios. The results show that our model outperforms the state-of-the-art
factorization and deep learning models in both accuracy and efficiency.
| no_new_dataset | 0.946101 |
1610.09300 | Antoine Gautier | Antoine Gautier, Quynh Nguyen and Matthias Hein | Globally Optimal Training of Generalized Polynomial Neural Networks with
Nonlinear Spectral Methods | Long version of NIPS 2016 paper | null | null | null | cs.LG math.OC stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The optimization problem behind neural networks is highly non-convex.
Training with stochastic gradient descent and variants requires careful
parameter tuning and provides no guarantee to achieve the global optimum. In
contrast we show under quite weak assumptions on the data that a particular
class of feedforward neural networks can be trained globally optimal with a
linear convergence rate with our nonlinear spectral method. Up to our knowledge
this is the first practically feasible method which achieves such a guarantee.
While the method can in principle be applied to deep networks, we restrict
ourselves for simplicity in this paper to one and two hidden layer networks.
Our experiments confirm that these models are rich enough to achieve good
performance on a series of real-world datasets.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2016 16:28:23 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Gautier",
"Antoine",
""
],
[
"Nguyen",
"Quynh",
""
],
[
"Hein",
"Matthias",
""
]
] | TITLE: Globally Optimal Training of Generalized Polynomial Neural Networks with
Nonlinear Spectral Methods
ABSTRACT: The optimization problem behind neural networks is highly non-convex.
Training with stochastic gradient descent and variants requires careful
parameter tuning and provides no guarantee to achieve the global optimum. In
contrast we show under quite weak assumptions on the data that a particular
class of feedforward neural networks can be trained globally optimal with a
linear convergence rate with our nonlinear spectral method. Up to our knowledge
this is the first practically feasible method which achieves such a guarantee.
While the method can in principle be applied to deep networks, we restrict
ourselves for simplicity in this paper to one and two hidden layer networks.
Our experiments confirm that these models are rich enough to achieve good
performance on a series of real-world datasets.
| no_new_dataset | 0.946498 |
1610.09334 | Seungryul Baek | Seungryul Baek, Kwang In Kim, Tae-Kyun Kim | Real-time Online Action Detection Forests using Spatio-temporal Contexts | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online action detection (OAD) is challenging since 1) robust yet
computationally expensive features cannot be straightforwardly used due to the
real-time processing requirements and 2) the localization and classification of
actions have to be performed even before they are fully observed. We propose a
new random forest (RF)-based online action detection framework that addresses
these challenges. Our algorithm uses computationally efficient skeletal joint
features. High accuracy is achieved by using robust convolutional neural
network (CNN)-based features which are extracted from the raw RGBD images, plus
the temporal relationships between the current frame of interest, and the past
and future frames. While these high-quality features are not available in
real-time testing scenario, we demonstrate that they can be effectively
exploited in training RF classifiers: We use these spatio-temporal contexts to
craft RF's new split functions improving RFs' leaf node statistics. Experiments
with challenging MSRAction3D, G3D, and OAD datasets demonstrate that our
algorithm significantly improves the accuracy over the state-of-the-art online
action detection algorithms while achieving the real-time efficiency of
existing skeleton-based RF classifiers.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2016 18:15:31 GMT"
}
] | 2016-10-31T00:00:00 | [
[
"Baek",
"Seungryul",
""
],
[
"Kim",
"Kwang In",
""
],
[
"Kim",
"Tae-Kyun",
""
]
] | TITLE: Real-time Online Action Detection Forests using Spatio-temporal Contexts
ABSTRACT: Online action detection (OAD) is challenging since 1) robust yet
computationally expensive features cannot be straightforwardly used due to the
real-time processing requirements and 2) the localization and classification of
actions have to be performed even before they are fully observed. We propose a
new random forest (RF)-based online action detection framework that addresses
these challenges. Our algorithm uses computationally efficient skeletal joint
features. High accuracy is achieved by using robust convolutional neural
network (CNN)-based features which are extracted from the raw RGBD images, plus
the temporal relationships between the current frame of interest, and the past
and future frames. While these high-quality features are not available in
real-time testing scenario, we demonstrate that they can be effectively
exploited in training RF classifiers: We use these spatio-temporal contexts to
craft RF's new split functions improving RFs' leaf node statistics. Experiments
with challenging MSRAction3D, G3D, and OAD datasets demonstrate that our
algorithm significantly improves the accuracy over the state-of-the-art online
action detection algorithms while achieving the real-time efficiency of
existing skeleton-based RF classifiers.
| no_new_dataset | 0.942981 |
1610.08851 | Andru Putra Twinanda | Andru P. Twinanda, Didier Mutter, Jacques Marescaux, Michel de
Mathelin, Nicolas Padoy | Single- and Multi-Task Architectures for Tool Presence Detection
Challenge at M2CAI 2016 | The dataset is available at http://camma.u-strasbg.fr/m2cai2016/ .
arXiv admin note: text overlap with arXiv:1610.08844 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The tool presence detection challenge at M2CAI 2016 consists of identifying
the presence/absence of seven surgical tools in the images of cholecystectomy
videos. Here, we propose to use deep architectures that are based on our
previous work where we presented several architectures to perform multiple
recognition tasks on laparoscopic videos. In this technical report, we present
the tool presence detection results using two architectures: (1) a single-task
architecture designed to perform solely the tool presence detection task and
(2) a multi-task architecture designed to perform jointly phase recognition and
tool presence detection. The results show that the multi-task network only
slightly improves the tool presence detection results. In constrast, a
significant improvement is obtained when there are more data available to train
the networks. This significant improvement can be regarded as a call for action
for other institutions to start working toward publishing more datasets into
the community, so that better models could be generated to perform the task.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 15:51:53 GMT"
}
] | 2016-10-30T00:00:00 | [
[
"Twinanda",
"Andru P.",
""
],
[
"Mutter",
"Didier",
""
],
[
"Marescaux",
"Jacques",
""
],
[
"de Mathelin",
"Michel",
""
],
[
"Padoy",
"Nicolas",
""
]
] | TITLE: Single- and Multi-Task Architectures for Tool Presence Detection
Challenge at M2CAI 2016
ABSTRACT: The tool presence detection challenge at M2CAI 2016 consists of identifying
the presence/absence of seven surgical tools in the images of cholecystectomy
videos. Here, we propose to use deep architectures that are based on our
previous work where we presented several architectures to perform multiple
recognition tasks on laparoscopic videos. In this technical report, we present
the tool presence detection results using two architectures: (1) a single-task
architecture designed to perform solely the tool presence detection task and
(2) a multi-task architecture designed to perform jointly phase recognition and
tool presence detection. The results show that the multi-task network only
slightly improves the tool presence detection results. In constrast, a
significant improvement is obtained when there are more data available to train
the networks. This significant improvement can be regarded as a call for action
for other institutions to start working toward publishing more datasets into
the community, so that better models could be generated to perform the task.
| no_new_dataset | 0.947235 |
1303.4778 | Eva Dyer | Eva L. Dyer, Aswin C. Sankaranarayanan, Richard G. Baraniuk | Greedy Feature Selection for Subspace Clustering | 32 pages, 7 figures, 1 table | Journal of Machine Learning Research, Vol.14, Issue 1, pp.
2487-2517, January 2013 | null | null | cs.LG math.NA stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unions of subspaces provide a powerful generalization to linear subspace
models for collections of high-dimensional data. To learn a union of subspaces
from a collection of data, sets of signals in the collection that belong to the
same subspace must be identified in order to obtain accurate estimates of the
subspace structures present in the data. Recently, sparse recovery methods have
been shown to provide a provable and robust strategy for exact feature
selection (EFS)--recovering subsets of points from the ensemble that live in
the same subspace. In parallel with recent studies of EFS with L1-minimization,
in this paper, we develop sufficient conditions for EFS with a greedy method
for sparse signal recovery known as orthogonal matching pursuit (OMP).
Following our analysis, we provide an empirical study of feature selection
strategies for signals living on unions of subspaces and characterize the gap
between sparse recovery methods and nearest neighbor (NN)-based approaches. In
particular, we demonstrate that sparse recovery methods provide significant
advantages over NN methods and the gap between the two approaches is
particularly pronounced when the sampling of subspaces in the dataset is
sparse. Our results suggest that OMP may be employed to reliably recover exact
feature sets in a number of regimes where NN approaches fail to reveal the
subspace membership of points in the ensemble.
| [
{
"version": "v1",
"created": "Tue, 19 Mar 2013 22:17:20 GMT"
},
{
"version": "v2",
"created": "Wed, 3 Jul 2013 19:07:34 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Dyer",
"Eva L.",
""
],
[
"Sankaranarayanan",
"Aswin C.",
""
],
[
"Baraniuk",
"Richard G.",
""
]
] | TITLE: Greedy Feature Selection for Subspace Clustering
ABSTRACT: Unions of subspaces provide a powerful generalization to linear subspace
models for collections of high-dimensional data. To learn a union of subspaces
from a collection of data, sets of signals in the collection that belong to the
same subspace must be identified in order to obtain accurate estimates of the
subspace structures present in the data. Recently, sparse recovery methods have
been shown to provide a provable and robust strategy for exact feature
selection (EFS)--recovering subsets of points from the ensemble that live in
the same subspace. In parallel with recent studies of EFS with L1-minimization,
in this paper, we develop sufficient conditions for EFS with a greedy method
for sparse signal recovery known as orthogonal matching pursuit (OMP).
Following our analysis, we provide an empirical study of feature selection
strategies for signals living on unions of subspaces and characterize the gap
between sparse recovery methods and nearest neighbor (NN)-based approaches. In
particular, we demonstrate that sparse recovery methods provide significant
advantages over NN methods and the gap between the two approaches is
particularly pronounced when the sampling of subspaces in the dataset is
sparse. Our results suggest that OMP may be employed to reliably recover exact
feature sets in a number of regimes where NN approaches fail to reveal the
subspace membership of points in the ensemble.
| no_new_dataset | 0.9455 |
1503.08169 | Azalia Mirhoseini | Azalia Mirhoseini, Eva L. Dyer, Ebrahim.M. Songhori, Richard G.
Baraniuk, Farinaz Koushanfar | RankMap: A Platform-Aware Framework for Distributed Learning from Dense
Datasets | 13 pages, 10 figures | null | null | null | cs.DC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces RankMap, a platform-aware end-to-end framework for
efficient execution of a broad class of iterative learning algorithms for
massive and dense datasets. Our framework exploits data structure to factorize
it into an ensemble of lower rank subspaces. The factorization creates sparse
low-dimensional representations of the data, a property which is leveraged to
devise effective mapping and scheduling of iterative learning algorithms on the
distributed computing machines. We provide two APIs, one matrix-based and one
graph-based, which facilitate automated adoption of the framework for
performing several contemporary learning applications. To demonstrate the
utility of RankMap, we solve sparse recovery and power iteration problems on
various real-world datasets with up to 1.8 billion non-zeros. Our evaluations
are performed on Amazon EC2 and IBM iDataPlex servers using up to 244 cores.
The results demonstrate up to two orders of magnitude improvements in memory
usage, execution speed, and bandwidth compared with the best reported prior
work, while achieving the same level of learning accuracy.
| [
{
"version": "v1",
"created": "Fri, 27 Mar 2015 18:02:51 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Oct 2016 14:29:44 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Mirhoseini",
"Azalia",
""
],
[
"Dyer",
"Eva L.",
""
],
[
"Songhori",
"Ebrahim. M.",
""
],
[
"Baraniuk",
"Richard G.",
""
],
[
"Koushanfar",
"Farinaz",
""
]
] | TITLE: RankMap: A Platform-Aware Framework for Distributed Learning from Dense
Datasets
ABSTRACT: This paper introduces RankMap, a platform-aware end-to-end framework for
efficient execution of a broad class of iterative learning algorithms for
massive and dense datasets. Our framework exploits data structure to factorize
it into an ensemble of lower rank subspaces. The factorization creates sparse
low-dimensional representations of the data, a property which is leveraged to
devise effective mapping and scheduling of iterative learning algorithms on the
distributed computing machines. We provide two APIs, one matrix-based and one
graph-based, which facilitate automated adoption of the framework for
performing several contemporary learning applications. To demonstrate the
utility of RankMap, we solve sparse recovery and power iteration problems on
various real-world datasets with up to 1.8 billion non-zeros. Our evaluations
are performed on Amazon EC2 and IBM iDataPlex servers using up to 244 cores.
The results demonstrate up to two orders of magnitude improvements in memory
usage, execution speed, and bandwidth compared with the best reported prior
work, while achieving the same level of learning accuracy.
| no_new_dataset | 0.946745 |
1505.00468 | Aishwarya Agrawal | Aishwarya Agrawal, Jiasen Lu, Stanislaw Antol, Margaret Mitchell, C.
Lawrence Zitnick, Dhruv Batra, Devi Parikh | VQA: Visual Question Answering | The first three authors contributed equally. International Conference
on Computer Vision (ICCV) 2015 | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose the task of free-form and open-ended Visual Question Answering
(VQA). Given an image and a natural language question about the image, the task
is to provide an accurate natural language answer. Mirroring real-world
scenarios, such as helping the visually impaired, both the questions and
answers are open-ended. Visual questions selectively target different areas of
an image, including background details and underlying context. As a result, a
system that succeeds at VQA typically needs a more detailed understanding of
the image and complex reasoning than a system producing generic image captions.
Moreover, VQA is amenable to automatic evaluation, since many open-ended
answers contain only a few words or a closed set of answers that can be
provided in a multiple-choice format. We provide a dataset containing ~0.25M
images, ~0.76M questions, and ~10M answers (www.visualqa.org), and discuss the
information it provides. Numerous baselines and methods for VQA are provided
and compared with human performance. Our VQA demo is available on CloudCV
(http://cloudcv.org/vqa).
| [
{
"version": "v1",
"created": "Sun, 3 May 2015 20:07:39 GMT"
},
{
"version": "v2",
"created": "Tue, 16 Jun 2015 16:59:52 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Oct 2015 02:47:20 GMT"
},
{
"version": "v4",
"created": "Wed, 18 Nov 2015 16:43:33 GMT"
},
{
"version": "v5",
"created": "Mon, 7 Mar 2016 20:55:28 GMT"
},
{
"version": "v6",
"created": "Wed, 20 Apr 2016 03:09:33 GMT"
},
{
"version": "v7",
"created": "Thu, 27 Oct 2016 03:50:19 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Agrawal",
"Aishwarya",
""
],
[
"Lu",
"Jiasen",
""
],
[
"Antol",
"Stanislaw",
""
],
[
"Mitchell",
"Margaret",
""
],
[
"Zitnick",
"C. Lawrence",
""
],
[
"Batra",
"Dhruv",
""
],
[
"Parikh",
"Devi",
""
]
] | TITLE: VQA: Visual Question Answering
ABSTRACT: We propose the task of free-form and open-ended Visual Question Answering
(VQA). Given an image and a natural language question about the image, the task
is to provide an accurate natural language answer. Mirroring real-world
scenarios, such as helping the visually impaired, both the questions and
answers are open-ended. Visual questions selectively target different areas of
an image, including background details and underlying context. As a result, a
system that succeeds at VQA typically needs a more detailed understanding of
the image and complex reasoning than a system producing generic image captions.
Moreover, VQA is amenable to automatic evaluation, since many open-ended
answers contain only a few words or a closed set of answers that can be
provided in a multiple-choice format. We provide a dataset containing ~0.25M
images, ~0.76M questions, and ~10M answers (www.visualqa.org), and discuss the
information it provides. Numerous baselines and methods for VQA are provided
and compared with human performance. Our VQA demo is available on CloudCV
(http://cloudcv.org/vqa).
| new_dataset | 0.957636 |
1506.04843 | Anirban Dasgupta | Anirban Dasgupta, Suvodip Chakrborty, Aritra Chaudhuri, Aurobinda
Routray | Evaluation of Denoising Techniques for EOG signals based on SNR
Estimation | in IEEE 2016 International Conference on Systems in Medicine and
Biology (ICSMB) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper evaluates four algorithms for denoising raw Electrooculography
(EOG) data based on the Signal to Noise Ratio (SNR). The SNR is computed using
the eigenvalue method. The filtering algorithms are a) Finite Impulse Response
(FIR) bandpass filters, b) Stationary Wavelet Transform, c) Empirical Mode
Decomposition (EMD) d) FIR Median Hybrid Filters. An EOG dataset has been
prepared where the subject is asked to perform letter cancelation test on 20
subjects.
| [
{
"version": "v1",
"created": "Tue, 16 Jun 2015 06:07:21 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Oct 2016 12:27:47 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Dasgupta",
"Anirban",
""
],
[
"Chakrborty",
"Suvodip",
""
],
[
"Chaudhuri",
"Aritra",
""
],
[
"Routray",
"Aurobinda",
""
]
] | TITLE: Evaluation of Denoising Techniques for EOG signals based on SNR
Estimation
ABSTRACT: This paper evaluates four algorithms for denoising raw Electrooculography
(EOG) data based on the Signal to Noise Ratio (SNR). The SNR is computed using
the eigenvalue method. The filtering algorithms are a) Finite Impulse Response
(FIR) bandpass filters, b) Stationary Wavelet Transform, c) Empirical Mode
Decomposition (EMD) d) FIR Median Hybrid Filters. An EOG dataset has been
prepared where the subject is asked to perform letter cancelation test on 20
subjects.
| new_dataset | 0.948251 |
1607.06657 | Ge Ou | Yan Wang, Ge Ou, Wei Pang, Lan Huang, George Macleod Coghill | e-Distance Weighted Support Vector Regression | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel support vector regression approach called e-Distance
Weighted Support Vector Regression (e-DWSVR).e-DWSVR specifically addresses two
challenging issues in support vector regression: first, the process of noisy
data; second, how to deal with the situation when the distribution of boundary
data is different from that of the overall data. The proposed e-DWSVR optimizes
the minimum margin and the mean of functional margin simultaneously to tackle
these two issues. In addition, we use both dual coordinate descent (CD) and
averaged stochastic gradient descent (ASGD) strategies to make e-DWSVR scalable
to large scale problems. We report promising results obtained by e-DWSVR in
comparison with existing methods on several benchmark datasets.
| [
{
"version": "v1",
"created": "Thu, 21 Jul 2016 02:35:57 GMT"
},
{
"version": "v2",
"created": "Tue, 9 Aug 2016 05:03:31 GMT"
},
{
"version": "v3",
"created": "Wed, 31 Aug 2016 08:28:10 GMT"
},
{
"version": "v4",
"created": "Thu, 27 Oct 2016 10:47:49 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Wang",
"Yan",
""
],
[
"Ou",
"Ge",
""
],
[
"Pang",
"Wei",
""
],
[
"Huang",
"Lan",
""
],
[
"Coghill",
"George Macleod",
""
]
] | TITLE: e-Distance Weighted Support Vector Regression
ABSTRACT: We propose a novel support vector regression approach called e-Distance
Weighted Support Vector Regression (e-DWSVR).e-DWSVR specifically addresses two
challenging issues in support vector regression: first, the process of noisy
data; second, how to deal with the situation when the distribution of boundary
data is different from that of the overall data. The proposed e-DWSVR optimizes
the minimum margin and the mean of functional margin simultaneously to tackle
these two issues. In addition, we use both dual coordinate descent (CD) and
averaged stochastic gradient descent (ASGD) strategies to make e-DWSVR scalable
to large scale problems. We report promising results obtained by e-DWSVR in
comparison with existing methods on several benchmark datasets.
| no_new_dataset | 0.949248 |
1610.01374 | Samik Banerjee | Samik Banerjee, Sukhendu Das | Domain Adaptation with Soft-margin multiple feature-kernel learning
beats Deep Learning for surveillance face recognition | This is an extended version of the paper accepted in CVPR Biometric
Workshop, 2016. arXiv admin note: text overlap with arXiv:1610.00660 | null | null | null | cs.CV cs.AI cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face recognition (FR) is the most preferred mode for biometric-based
surveillance, due to its passive nature of detecting subjects, amongst all
different types of biometric traits. FR under surveillance scenario does not
give satisfactory performance due to low contrast, noise and poor illumination
conditions on probes, as compared to the training samples. A state-of-the-art
technology, Deep Learning, even fails to perform well in these scenarios. We
propose a novel soft-margin based learning method for multiple feature-kernel
combinations, followed by feature transformed using Domain Adaptation, which
outperforms many recent state-of-the-art techniques, when tested using three
real-world surveillance face datasets.
| [
{
"version": "v1",
"created": "Wed, 5 Oct 2016 11:48:56 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Oct 2016 13:14:49 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Banerjee",
"Samik",
""
],
[
"Das",
"Sukhendu",
""
]
] | TITLE: Domain Adaptation with Soft-margin multiple feature-kernel learning
beats Deep Learning for surveillance face recognition
ABSTRACT: Face recognition (FR) is the most preferred mode for biometric-based
surveillance, due to its passive nature of detecting subjects, amongst all
different types of biometric traits. FR under surveillance scenario does not
give satisfactory performance due to low contrast, noise and poor illumination
conditions on probes, as compared to the training samples. A state-of-the-art
technology, Deep Learning, even fails to perform well in these scenarios. We
propose a novel soft-margin based learning method for multiple feature-kernel
combinations, followed by feature transformed using Domain Adaptation, which
outperforms many recent state-of-the-art techniques, when tested using three
real-world surveillance face datasets.
| no_new_dataset | 0.94887 |
1610.08559 | Ke Yang | Ke Yang and Julia Stoyanovich | Measuring Fairness in Ranked Outputs | 5 pages, 7 figures, FATML 2016 | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ranking and scoring are ubiquitous. We consider the setting in which an
institution, called a ranker, evaluates a set of individuals based on
demographic, behavioral or other characteristics. The final output is a ranking
that represents the relative quality of the individuals. While automatic and
therefore seemingly objective, rankers can, and often do, discriminate against
individuals and systematically disadvantage members of protected groups. This
warrants a careful study of the fairness of a ranking scheme.
In this paper we propose fairness measures for ranked outputs. We develop a
data generation procedure that allows us to systematically control the degree
of unfairness in the output, and study the behavior of our measures on these
datasets. We then apply our proposed measures to several real datasets, and
demonstrate cases of unfairness. Finally, we show preliminary results of
incorporating our ranked fairness measures into an optimization framework, and
show potential for improving fairness of ranked outputs while maintaining
accuracy.
| [
{
"version": "v1",
"created": "Wed, 26 Oct 2016 22:02:39 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Yang",
"Ke",
""
],
[
"Stoyanovich",
"Julia",
""
]
] | TITLE: Measuring Fairness in Ranked Outputs
ABSTRACT: Ranking and scoring are ubiquitous. We consider the setting in which an
institution, called a ranker, evaluates a set of individuals based on
demographic, behavioral or other characteristics. The final output is a ranking
that represents the relative quality of the individuals. While automatic and
therefore seemingly objective, rankers can, and often do, discriminate against
individuals and systematically disadvantage members of protected groups. This
warrants a careful study of the fairness of a ranking scheme.
In this paper we propose fairness measures for ranked outputs. We develop a
data generation procedure that allows us to systematically control the degree
of unfairness in the output, and study the behavior of our measures on these
datasets. We then apply our proposed measures to several real datasets, and
demonstrate cases of unfairness. Finally, we show preliminary results of
incorporating our ranked fairness measures into an optimization framework, and
show potential for improving fairness of ranked outputs while maintaining
accuracy.
| no_new_dataset | 0.946892 |
1610.08624 | PeiXin Hou | Peixin Hou, Hao Deng, Jiguang Yue, and Shuguang Liu | PCM and APCM Revisited: An Uncertainty Perspective | 8 pages | null | null | null | cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we take a new look at the possibilistic c-means (PCM) and
adaptive PCM (APCM) clustering algorithms from the perspective of uncertainty.
This new perspective offers us insights into the clustering process, and also
provides us greater degree of flexibility. We analyze the clustering behavior
of PCM-based algorithms and introduce parameters $\sigma_v$ and $\alpha$ to
characterize uncertainty of estimated bandwidth and noise level of the dataset
respectively. Then uncertainty (fuzziness) of membership values caused by
uncertainty of the estimated bandwidth parameter is modeled by a conditional
fuzzy set, which is a new formulation of the type-2 fuzzy set. Experiments show
that parameters $\sigma_v$ and $\alpha$ make the clustering process more easy
to control, and main features of PCM and APCM are unified in this new
clustering framework (UPCM). More specifically, UPCM reduces to PCM when we set
a small $\alpha$ or a large $\sigma_v$, and UPCM reduces to APCM when clusters
are confined in their physical clusters and possible cluster elimination are
ensured. Finally we present further researches of this paper.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 05:41:23 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Hou",
"Peixin",
""
],
[
"Deng",
"Hao",
""
],
[
"Yue",
"Jiguang",
""
],
[
"Liu",
"Shuguang",
""
]
] | TITLE: PCM and APCM Revisited: An Uncertainty Perspective
ABSTRACT: In this paper, we take a new look at the possibilistic c-means (PCM) and
adaptive PCM (APCM) clustering algorithms from the perspective of uncertainty.
This new perspective offers us insights into the clustering process, and also
provides us greater degree of flexibility. We analyze the clustering behavior
of PCM-based algorithms and introduce parameters $\sigma_v$ and $\alpha$ to
characterize uncertainty of estimated bandwidth and noise level of the dataset
respectively. Then uncertainty (fuzziness) of membership values caused by
uncertainty of the estimated bandwidth parameter is modeled by a conditional
fuzzy set, which is a new formulation of the type-2 fuzzy set. Experiments show
that parameters $\sigma_v$ and $\alpha$ make the clustering process more easy
to control, and main features of PCM and APCM are unified in this new
clustering framework (UPCM). More specifically, UPCM reduces to PCM when we set
a small $\alpha$ or a large $\sigma_v$, and UPCM reduces to APCM when clusters
are confined in their physical clusters and possible cluster elimination are
ensured. Finally we present further researches of this paper.
| no_new_dataset | 0.9455 |
1610.08640 | Marc Schoenauer | Marti Luis (TAO, LRI), Fansi-Tchango Arsene (TRT), Navarro Laurent
(TRT), Marc Schoenauer (TAO, LRI) | Anomaly Detection with the Voronoi Diagram Evolutionary Algorithm | null | Parallel Problem Solving from Nature -- PPSN XIV, Sep 2016,
Edinburgh, France. Springer Verlag, 9921, pp.697-706, 2016, LNCS | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the Voronoi diagram-based evolutionary algorithm
(VorEAl). VorEAl partitions input space in abnormal/normal subsets using
Voronoi diagrams. Diagrams are evolved using a multi-objective bio-inspired
approach in order to conjointly optimize classification metrics while also
being able to represent areas of the data space that are not present in the
training dataset. As part of the paper VorEAl is experimentally validated and
contrasted with similar approaches.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 07:05:54 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Luis",
"Marti",
"",
"TAO, LRI"
],
[
"Arsene",
"Fansi-Tchango",
"",
"TRT"
],
[
"Laurent",
"Navarro",
"",
"TRT"
],
[
"Schoenauer",
"Marc",
"",
"TAO, LRI"
]
] | TITLE: Anomaly Detection with the Voronoi Diagram Evolutionary Algorithm
ABSTRACT: This paper presents the Voronoi diagram-based evolutionary algorithm
(VorEAl). VorEAl partitions input space in abnormal/normal subsets using
Voronoi diagrams. Diagrams are evolved using a multi-objective bio-inspired
approach in order to conjointly optimize classification metrics while also
being able to represent areas of the data space that are not present in the
training dataset. As part of the paper VorEAl is experimentally validated and
contrasted with similar approaches.
| no_new_dataset | 0.948251 |
1610.08686 | Mauro Coletto | Mauro Coletto, Claudio Lucchese, Salvatore Orlando, Raffaele Perego | Polarized User and Topic Tracking in Twitter | SIGIR 16 | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Digital traces of conversations in micro-blogging platforms and OSNs provide
information about user opinion with a high degree of resolution. These
information sources can be exploited to under- stand and monitor collective
behaviors. In this work, we focus on polarization classes, i.e., those topics
that require the user to side exclusively with one position. The proposed
method provides an iterative classification of users and keywords: first,
polarized users are identified, then polarized keywords are discovered by
monitoring the activities of previously classified users. This method thus
allows tracking users and topics over time. We report several experiments
conducted on two Twitter datasets during political election time-frames. We
measure the user classification accuracy on a golden set of users, and analyze
the relevance of the extracted keywords for the ongoing political discussion.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 10:03:31 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Coletto",
"Mauro",
""
],
[
"Lucchese",
"Claudio",
""
],
[
"Orlando",
"Salvatore",
""
],
[
"Perego",
"Raffaele",
""
]
] | TITLE: Polarized User and Topic Tracking in Twitter
ABSTRACT: Digital traces of conversations in micro-blogging platforms and OSNs provide
information about user opinion with a high degree of resolution. These
information sources can be exploited to under- stand and monitor collective
behaviors. In this work, we focus on polarization classes, i.e., those topics
that require the user to side exclusively with one position. The proposed
method provides an iterative classification of users and keywords: first,
polarized users are identified, then polarized keywords are discovered by
monitoring the activities of previously classified users. This method thus
allows tracking users and topics over time. We report several experiments
conducted on two Twitter datasets during political election time-frames. We
measure the user classification accuracy on a golden set of users, and analyze
the relevance of the extracted keywords for the ongoing political discussion.
| no_new_dataset | 0.951006 |
1610.08739 | Andre Droschinsky | Andre Droschinsky and Nils Kriege and Petra Mutzel | Finding Largest Common Substructures of Molecules in Quadratic Time | null | null | null | null | cs.DS cs.CC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Finding the common structural features of two molecules is a fundamental task
in cheminformatics. Most drugs are small molecules, which can naturally be
interpreted as graphs. Hence, the task is formalized as maximum common subgraph
problem. Albeit the vast majority of molecules yields outerplanar graphs this
problem remains NP-hard.
We consider a variation of the problem of high practical relevance, where the
rings of molecules must not be broken, i.e., the block and bridge structure of
the input graphs must be retained by the common subgraph. We present an
algorithm for finding a maximum common connected induced subgraph of two given
outerplanar graphs subject to this constraint. Our approach runs in time
$\mathcal{O}(\Delta n^2)$ in outerplanar graphs on $n$ vertices with maximum
degree $\Delta$. This leads to a quadratic time complexity in molecular graphs,
which have bounded degree. The experimental comparison on synthetic and
real-world datasets shows that our approach is highly efficient in practice and
outperforms comparable state-of-the-art algorithms.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 12:16:01 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Droschinsky",
"Andre",
""
],
[
"Kriege",
"Nils",
""
],
[
"Mutzel",
"Petra",
""
]
] | TITLE: Finding Largest Common Substructures of Molecules in Quadratic Time
ABSTRACT: Finding the common structural features of two molecules is a fundamental task
in cheminformatics. Most drugs are small molecules, which can naturally be
interpreted as graphs. Hence, the task is formalized as maximum common subgraph
problem. Albeit the vast majority of molecules yields outerplanar graphs this
problem remains NP-hard.
We consider a variation of the problem of high practical relevance, where the
rings of molecules must not be broken, i.e., the block and bridge structure of
the input graphs must be retained by the common subgraph. We present an
algorithm for finding a maximum common connected induced subgraph of two given
outerplanar graphs subject to this constraint. Our approach runs in time
$\mathcal{O}(\Delta n^2)$ in outerplanar graphs on $n$ vertices with maximum
degree $\Delta$. This leads to a quadratic time complexity in molecular graphs,
which have bounded degree. The experimental comparison on synthetic and
real-world datasets shows that our approach is highly efficient in practice and
outperforms comparable state-of-the-art algorithms.
| no_new_dataset | 0.945045 |
1610.08854 | Manish Sahu | Manish Sahu, Anirban Mukhopadhyay, Angelika Szengel and Stefan Zachow | Tool and Phase recognition using contextual CNN features | MICCAI M2CAI 2016 Surgical tool & phase detection challenge report | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A transfer learning method for generating features suitable for surgical
tools and phase recognition from the ImageNet classification features [1] is
proposed here. In addition, methods are developed for generating contextual
features and combining them with time series analysis for final classification
using multi-class random forest. The proposed pipeline is tested over the
training and testing datasets of M2CAI16 challenges: tool and phase detection.
Encouraging results are obtained by leave-one-out cross validation evaluation
on the training dataset.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 15:54:41 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Sahu",
"Manish",
""
],
[
"Mukhopadhyay",
"Anirban",
""
],
[
"Szengel",
"Angelika",
""
],
[
"Zachow",
"Stefan",
""
]
] | TITLE: Tool and Phase recognition using contextual CNN features
ABSTRACT: A transfer learning method for generating features suitable for surgical
tools and phase recognition from the ImageNet classification features [1] is
proposed here. In addition, methods are developed for generating contextual
features and combining them with time series analysis for final classification
using multi-class random forest. The proposed pipeline is tested over the
training and testing datasets of M2CAI16 challenges: tool and phase detection.
Encouraging results are obtained by leave-one-out cross validation evaluation
on the training dataset.
| no_new_dataset | 0.945147 |
1610.08871 | Nicholas Westlake | Nicholas Westlake, Hongping Cai and Peter Hall | Detecting People in Artwork with CNNs | 14 pages, plus 3 pages of references; 7 figures in ECCV 2016
Workshops | null | 10.1007/978-3-319-46604-0_57 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | CNNs have massively improved performance in object detection in photographs.
However research into object detection in artwork remains limited. We show
state-of-the-art performance on a challenging dataset, People-Art, which
contains people from photos, cartoons and 41 different artwork movements. We
achieve this high performance by fine-tuning a CNN for this task, thus also
demonstrating that training CNNs on photos results in overfitting for photos:
only the first three or four layers transfer from photos to artwork. Although
the CNN's performance is the highest yet, it remains less than 60\% AP,
suggesting further work is needed for the cross-depiction problem. The final
publication is available at Springer via
http://dx.doi.org/10.1007/978-3-319-46604-0_57
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 16:30:15 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Westlake",
"Nicholas",
""
],
[
"Cai",
"Hongping",
""
],
[
"Hall",
"Peter",
""
]
] | TITLE: Detecting People in Artwork with CNNs
ABSTRACT: CNNs have massively improved performance in object detection in photographs.
However research into object detection in artwork remains limited. We show
state-of-the-art performance on a challenging dataset, People-Art, which
contains people from photos, cartoons and 41 different artwork movements. We
achieve this high performance by fine-tuning a CNN for this task, thus also
demonstrating that training CNNs on photos results in overfitting for photos:
only the first three or four layers transfer from photos to artwork. Although
the CNN's performance is the highest yet, it remains less than 60\% AP,
suggesting further work is needed for the cross-depiction problem. The final
publication is available at Springer via
http://dx.doi.org/10.1007/978-3-319-46604-0_57
| no_new_dataset | 0.933491 |
1610.08904 | Chen Huang | Chen Huang, Chen Change Loy, Xiaoou Tang | Local Similarity-Aware Deep Feature Embedding | 9 pages, 4 figures, 2 tables. Accepted to NIPS 2016 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing deep embedding methods in vision tasks are capable of learning a
compact Euclidean space from images, where Euclidean distances correspond to a
similarity metric. To make learning more effective and efficient, hard sample
mining is usually employed, with samples identified through computing the
Euclidean feature distance. However, the global Euclidean distance cannot
faithfully characterize the true feature similarity in a complex visual feature
space, where the intraclass distance in a high-density region may be larger
than the interclass distance in low-density regions. In this paper, we
introduce a Position-Dependent Deep Metric (PDDM) unit, which is capable of
learning a similarity metric adaptive to local feature structure. The metric
can be used to select genuinely hard samples in a local neighborhood to guide
the deep embedding learning in an online and robust manner. The new layer is
appealing in that it is pluggable to any convolutional networks and is trained
end-to-end. Our local similarity-aware feature embedding not only demonstrates
faster convergence and boosted performance on two complex image retrieval
datasets, its large margin nature also leads to superior generalization results
under the large and open set scenarios of transfer learning and zero-shot
learning on ImageNet 2010 and ImageNet-10K datasets.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2016 17:51:18 GMT"
}
] | 2016-10-28T00:00:00 | [
[
"Huang",
"Chen",
""
],
[
"Loy",
"Chen Change",
""
],
[
"Tang",
"Xiaoou",
""
]
] | TITLE: Local Similarity-Aware Deep Feature Embedding
ABSTRACT: Existing deep embedding methods in vision tasks are capable of learning a
compact Euclidean space from images, where Euclidean distances correspond to a
similarity metric. To make learning more effective and efficient, hard sample
mining is usually employed, with samples identified through computing the
Euclidean feature distance. However, the global Euclidean distance cannot
faithfully characterize the true feature similarity in a complex visual feature
space, where the intraclass distance in a high-density region may be larger
than the interclass distance in low-density regions. In this paper, we
introduce a Position-Dependent Deep Metric (PDDM) unit, which is capable of
learning a similarity metric adaptive to local feature structure. The metric
can be used to select genuinely hard samples in a local neighborhood to guide
the deep embedding learning in an online and robust manner. The new layer is
appealing in that it is pluggable to any convolutional networks and is trained
end-to-end. Our local similarity-aware feature embedding not only demonstrates
faster convergence and boosted performance on two complex image retrieval
datasets, its large margin nature also leads to superior generalization results
under the large and open set scenarios of transfer learning and zero-shot
learning on ImageNet 2010 and ImageNet-10K datasets.
| no_new_dataset | 0.948298 |
1608.05995 | Ming Lin | Ming Lin and Jieping Ye | A Non-convex One-Pass Framework for Generalized Factorization Machine
and Rank-One Matrix Sensing | accepted by NIPS 2016 | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We develop an efficient alternating framework for learning a generalized
version of Factorization Machine (gFM) on steaming data with provable
guarantees. When the instances are sampled from $d$ dimensional random Gaussian
vectors and the target second order coefficient matrix in gFM is of rank $k$,
our algorithm converges linearly, achieves $O(\epsilon)$ recovery error after
retrieving $O(k^{3}d\log(1/\epsilon))$ training instances, consumes $O(kd)$
memory in one-pass of dataset and only requires matrix-vector product
operations in each iteration. The key ingredient of our framework is a
construction of an estimation sequence endowed with a so-called Conditionally
Independent RIP condition (CI-RIP). As special cases of gFM, our framework can
be applied to symmetric or asymmetric rank-one matrix sensing problems, such as
inductive matrix completion and phase retrieval.
| [
{
"version": "v1",
"created": "Sun, 21 Aug 2016 20:28:29 GMT"
},
{
"version": "v2",
"created": "Fri, 9 Sep 2016 17:54:50 GMT"
},
{
"version": "v3",
"created": "Mon, 12 Sep 2016 21:43:05 GMT"
},
{
"version": "v4",
"created": "Wed, 14 Sep 2016 02:24:22 GMT"
},
{
"version": "v5",
"created": "Tue, 25 Oct 2016 21:23:23 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Lin",
"Ming",
""
],
[
"Ye",
"Jieping",
""
]
] | TITLE: A Non-convex One-Pass Framework for Generalized Factorization Machine
and Rank-One Matrix Sensing
ABSTRACT: We develop an efficient alternating framework for learning a generalized
version of Factorization Machine (gFM) on steaming data with provable
guarantees. When the instances are sampled from $d$ dimensional random Gaussian
vectors and the target second order coefficient matrix in gFM is of rank $k$,
our algorithm converges linearly, achieves $O(\epsilon)$ recovery error after
retrieving $O(k^{3}d\log(1/\epsilon))$ training instances, consumes $O(kd)$
memory in one-pass of dataset and only requires matrix-vector product
operations in each iteration. The key ingredient of our framework is a
construction of an estimation sequence endowed with a so-called Conditionally
Independent RIP condition (CI-RIP). As special cases of gFM, our framework can
be applied to symmetric or asymmetric rank-one matrix sensing problems, such as
inductive matrix completion and phase retrieval.
| no_new_dataset | 0.942823 |
1610.06656 | Shanshan Wu | Shanshan Wu, Srinadh Bhojanapalli, Sujay Sanghavi, Alexandros G.
Dimakis | Single Pass PCA of Matrix Products | 24 pages, 4 figures, NIPS 2016 | null | null | null | stat.ML cs.DS cs.IT cs.LG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we present a new algorithm for computing a low rank
approximation of the product $A^TB$ by taking only a single pass of the two
matrices $A$ and $B$. The straightforward way to do this is to (a) first sketch
$A$ and $B$ individually, and then (b) find the top components using PCA on the
sketch. Our algorithm in contrast retains additional summary information about
$A,B$ (e.g. row and column norms etc.) and uses this additional information to
obtain an improved approximation from the sketches. Our main analytical result
establishes a comparable spectral norm guarantee to existing two-pass methods;
in addition we also provide results from an Apache Spark implementation that
shows better computational and statistical performance on real-world and
synthetic evaluation datasets.
| [
{
"version": "v1",
"created": "Fri, 21 Oct 2016 02:45:46 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Oct 2016 13:58:24 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Wu",
"Shanshan",
""
],
[
"Bhojanapalli",
"Srinadh",
""
],
[
"Sanghavi",
"Sujay",
""
],
[
"Dimakis",
"Alexandros G.",
""
]
] | TITLE: Single Pass PCA of Matrix Products
ABSTRACT: In this paper we present a new algorithm for computing a low rank
approximation of the product $A^TB$ by taking only a single pass of the two
matrices $A$ and $B$. The straightforward way to do this is to (a) first sketch
$A$ and $B$ individually, and then (b) find the top components using PCA on the
sketch. Our algorithm in contrast retains additional summary information about
$A,B$ (e.g. row and column norms etc.) and uses this additional information to
obtain an improved approximation from the sketches. Our main analytical result
establishes a comparable spectral norm guarantee to existing two-pass methods;
in addition we also provide results from an Apache Spark implementation that
shows better computational and statistical performance on real-world and
synthetic evaluation datasets.
| no_new_dataset | 0.944074 |
1610.07667 | Nir Rosenfeld | Nir Rosenfeld, Yishay Mansour, Elad Yom-Tov | Predicting Counterfactuals from Large Historical Data and Small
Randomized Trials | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When a new treatment is considered for use, whether a pharmaceutical drug or
a search engine ranking algorithm, a typical question that arises is, will its
performance exceed that of the current treatment? The conventional way to
answer this counterfactual question is to estimate the effect of the new
treatment in comparison to that of the conventional treatment by running a
controlled, randomized experiment. While this approach theoretically ensures an
unbiased estimator, it suffers from several drawbacks, including the difficulty
in finding representative experimental populations as well as the cost of
running such trials. Moreover, such trials neglect the huge quantities of
available control-condition data which are often completely ignored.
In this paper we propose a discriminative framework for estimating the
performance of a new treatment given a large dataset of the control condition
and data from a small (and possibly unrepresentative) randomized trial
comparing new and old treatments. Our objective, which requires minimal
assumptions on the treatments, models the relation between the outcomes of the
different conditions. This allows us to not only estimate mean effects but also
to generate individual predictions for examples outside the randomized sample.
We demonstrate the utility of our approach through experiments in three
areas: Search engine operation, treatments to diabetes patients, and market
value estimation for houses. Our results demonstrate that our approach can
reduce the number and size of the currently performed randomized controlled
experiments, thus saving significant time, money and effort on the part of
practitioners.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 22:12:52 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Oct 2016 06:11:07 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Rosenfeld",
"Nir",
""
],
[
"Mansour",
"Yishay",
""
],
[
"Yom-Tov",
"Elad",
""
]
] | TITLE: Predicting Counterfactuals from Large Historical Data and Small
Randomized Trials
ABSTRACT: When a new treatment is considered for use, whether a pharmaceutical drug or
a search engine ranking algorithm, a typical question that arises is, will its
performance exceed that of the current treatment? The conventional way to
answer this counterfactual question is to estimate the effect of the new
treatment in comparison to that of the conventional treatment by running a
controlled, randomized experiment. While this approach theoretically ensures an
unbiased estimator, it suffers from several drawbacks, including the difficulty
in finding representative experimental populations as well as the cost of
running such trials. Moreover, such trials neglect the huge quantities of
available control-condition data which are often completely ignored.
In this paper we propose a discriminative framework for estimating the
performance of a new treatment given a large dataset of the control condition
and data from a small (and possibly unrepresentative) randomized trial
comparing new and old treatments. Our objective, which requires minimal
assumptions on the treatments, models the relation between the outcomes of the
different conditions. This allows us to not only estimate mean effects but also
to generate individual predictions for examples outside the randomized sample.
We demonstrate the utility of our approach through experiments in three
areas: Search engine operation, treatments to diabetes patients, and market
value estimation for houses. Our results demonstrate that our approach can
reduce the number and size of the currently performed randomized controlled
experiments, thus saving significant time, money and effort on the part of
practitioners.
| no_new_dataset | 0.94474 |
1610.08077 | James Johndrow | Kristian Lum and James Johndrow | A statistical framework for fair predictive algorithms | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predictive modeling is increasingly being employed to assist human
decision-makers. One purported advantage of replacing human judgment with
computer models in high stakes settings-- such as sentencing, hiring, policing,
college admissions, and parole decisions-- is the perceived "neutrality" of
computers. It is argued that because computer models do not hold personal
prejudice, the predictions they produce will be equally free from prejudice.
There is growing recognition that employing algorithms does not remove the
potential for bias, and can even amplify it, since training data were
inevitably generated by a process that is itself biased. In this paper, we
provide a probabilistic definition of algorithmic bias. We propose a method to
remove bias from predictive models by removing all information regarding
protected variables from the permitted training data. Unlike previous work in
this area, our framework is general enough to accommodate arbitrary data types,
e.g. binary, continuous, etc. Motivated by models currently in use in the
criminal justice system that inform decisions on pre-trial release and
paroling, we apply our proposed method to a dataset on the criminal histories
of individuals at the time of sentencing to produce "race-neutral" predictions
of re-arrest. In the process, we demonstrate that the most common approach to
creating "race-neutral" models-- omitting race as a covariate-- still results
in racially disparate predictions. We then demonstrate that the application of
our proposed method to these data removes racial disparities from predictions
with minimal impact on predictive accuracy.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 20:18:24 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Lum",
"Kristian",
""
],
[
"Johndrow",
"James",
""
]
] | TITLE: A statistical framework for fair predictive algorithms
ABSTRACT: Predictive modeling is increasingly being employed to assist human
decision-makers. One purported advantage of replacing human judgment with
computer models in high stakes settings-- such as sentencing, hiring, policing,
college admissions, and parole decisions-- is the perceived "neutrality" of
computers. It is argued that because computer models do not hold personal
prejudice, the predictions they produce will be equally free from prejudice.
There is growing recognition that employing algorithms does not remove the
potential for bias, and can even amplify it, since training data were
inevitably generated by a process that is itself biased. In this paper, we
provide a probabilistic definition of algorithmic bias. We propose a method to
remove bias from predictive models by removing all information regarding
protected variables from the permitted training data. Unlike previous work in
this area, our framework is general enough to accommodate arbitrary data types,
e.g. binary, continuous, etc. Motivated by models currently in use in the
criminal justice system that inform decisions on pre-trial release and
paroling, we apply our proposed method to a dataset on the criminal histories
of individuals at the time of sentencing to produce "race-neutral" predictions
of re-arrest. In the process, we demonstrate that the most common approach to
creating "race-neutral" models-- omitting race as a covariate-- still results
in racially disparate predictions. We then demonstrate that the application of
our proposed method to these data removes racial disparities from predictions
with minimal impact on predictive accuracy.
| no_new_dataset | 0.943138 |
1610.08095 | Mengting Wan | Mengting Wan, Julian McAuley | Modeling Ambiguity, Subjectivity, and Diverging Viewpoints in Opinion
Question Answering Systems | 10 pages, accepted by ICDM'2016 | null | null | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Product review websites provide an incredible lens into the wide variety of
opinions and experiences of different people, and play a critical role in
helping users discover products that match their personal needs and
preferences. To help address questions that can't easily be answered by reading
others' reviews, some review websites also allow users to pose questions to the
community via a question-answering (QA) system. As one would expect, just as
opinions diverge among different reviewers, answers to such questions may also
be subjective, opinionated, and divergent. This means that answering such
questions automatically is quite different from traditional QA tasks, where it
is assumed that a single `correct' answer is available. While recent work
introduced the idea of question-answering using product reviews, it did not
account for two aspects that we consider in this paper: (1) Questions have
multiple, often divergent, answers, and this full spectrum of answers should
somehow be used to train the system; and (2) What makes a `good' answer depends
on the asker and the answerer, and these factors should be incorporated in
order for the system to be more personalized. Here we build a new QA dataset
with 800 thousand questions---and over 3.1 million answers---and show that
explicitly accounting for personalization and ambiguity leads both to
quantitatively better answers, but also a more nuanced view of the range of
supporting, but subjective, opinions.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 21:08:15 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Wan",
"Mengting",
""
],
[
"McAuley",
"Julian",
""
]
] | TITLE: Modeling Ambiguity, Subjectivity, and Diverging Viewpoints in Opinion
Question Answering Systems
ABSTRACT: Product review websites provide an incredible lens into the wide variety of
opinions and experiences of different people, and play a critical role in
helping users discover products that match their personal needs and
preferences. To help address questions that can't easily be answered by reading
others' reviews, some review websites also allow users to pose questions to the
community via a question-answering (QA) system. As one would expect, just as
opinions diverge among different reviewers, answers to such questions may also
be subjective, opinionated, and divergent. This means that answering such
questions automatically is quite different from traditional QA tasks, where it
is assumed that a single `correct' answer is available. While recent work
introduced the idea of question-answering using product reviews, it did not
account for two aspects that we consider in this paper: (1) Questions have
multiple, often divergent, answers, and this full spectrum of answers should
somehow be used to train the system; and (2) What makes a `good' answer depends
on the asker and the answerer, and these factors should be incorporated in
order for the system to be more personalized. Here we build a new QA dataset
with 800 thousand questions---and over 3.1 million answers---and show that
explicitly accounting for personalization and ambiguity leads both to
quantitatively better answers, but also a more nuanced view of the range of
supporting, but subjective, opinions.
| new_dataset | 0.961207 |
1610.08120 | Suchet Bargoti | Suchet Bargoti, James Underwood | Image Segmentation for Fruit Detection and Yield Estimation in Apple
Orchards | This paper is the initial version of the manuscript submitted to The
Journal of Field Robotics in May 2016. Following reviews and revisions, the
paper has been accepted for publication. The reviewed version includes
extended comparison between the different classification frameworks and a
more in-depth literature review | null | null | null | cs.RO cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ground vehicles equipped with monocular vision systems are a valuable source
of high resolution image data for precision agriculture applications in
orchards. This paper presents an image processing framework for fruit detection
and counting using orchard image data. A general purpose image segmentation
approach is used, including two feature learning algorithms; multi-scale
Multi-Layered Perceptrons (MLP) and Convolutional Neural Networks (CNN). These
networks were extended by including contextual information about how the image
data was captured (metadata), which correlates with some of the appearance
variations and/or class distributions observed in the data. The pixel-wise
fruit segmentation output is processed using the Watershed Segmentation (WS)
and Circular Hough Transform (CHT) algorithms to detect and count individual
fruits. Experiments were conducted in a commercial apple orchard near
Melbourne, Australia. The results show an improvement in fruit segmentation
performance with the inclusion of metadata on the previously benchmarked MLP
network. We extend this work with CNNs, bringing agrovision closer to the
state-of-the-art in computer vision, where although metadata had negligible
influence, the best pixel-wise F1-score of $0.791$ was achieved. The WS
algorithm produced the best apple detection and counting results, with a
detection F1-score of $0.858$. As a final step, image fruit counts were
accumulated over multiple rows at the orchard and compared against the
post-harvest fruit counts that were obtained from a grading and counting
machine. The count estimates using CNN and WS resulted in the best performance
for this dataset, with a squared correlation coefficient of $r^2=0.826$.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 23:38:02 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Bargoti",
"Suchet",
""
],
[
"Underwood",
"James",
""
]
] | TITLE: Image Segmentation for Fruit Detection and Yield Estimation in Apple
Orchards
ABSTRACT: Ground vehicles equipped with monocular vision systems are a valuable source
of high resolution image data for precision agriculture applications in
orchards. This paper presents an image processing framework for fruit detection
and counting using orchard image data. A general purpose image segmentation
approach is used, including two feature learning algorithms; multi-scale
Multi-Layered Perceptrons (MLP) and Convolutional Neural Networks (CNN). These
networks were extended by including contextual information about how the image
data was captured (metadata), which correlates with some of the appearance
variations and/or class distributions observed in the data. The pixel-wise
fruit segmentation output is processed using the Watershed Segmentation (WS)
and Circular Hough Transform (CHT) algorithms to detect and count individual
fruits. Experiments were conducted in a commercial apple orchard near
Melbourne, Australia. The results show an improvement in fruit segmentation
performance with the inclusion of metadata on the previously benchmarked MLP
network. We extend this work with CNNs, bringing agrovision closer to the
state-of-the-art in computer vision, where although metadata had negligible
influence, the best pixel-wise F1-score of $0.791$ was achieved. The WS
algorithm produced the best apple detection and counting results, with a
detection F1-score of $0.858$. As a final step, image fruit counts were
accumulated over multiple rows at the orchard and compared against the
post-harvest fruit counts that were obtained from a grading and counting
machine. The count estimates using CNN and WS resulted in the best performance
for this dataset, with a squared correlation coefficient of $r^2=0.826$.
| no_new_dataset | 0.954478 |
1610.08133 | Elaheh Raisi | Hamid Abrishami Moghaddam and Elaheh Raisi | Incremental Nonparametric Weighted Feature Extraction for OnlineSubspace
Pattern Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, a new online method based on nonparametric weighted feature
extraction (NWFE) is proposed. NWFE was introduced to enjoy optimum
characteristics of linear discriminant analysis (LDA) and nonparametric
discriminant analysis (NDA) while rectifying their drawbacks. It emphasizes the
points near decision boundary by putting greater weights on them and
deemphasizes other points. Incremental nonparametric weighted feature
extraction (INWFE) is the online version of NWFE. INWFE has advantages of NWFE
method such as extracting more than L-1 features in contrast to LDA. It is
independent of the class distribution and performs well in complex distributed
data. The effects of outliers are reduced due to the nature of its
nonparametric scatter matrix. Furthermore, it is possible to add new samples
asynchronously, i.e. whenever a new sample becomes available at any given time,
it can be added to the algorithm. This is useful for many real world
applications since all data cannot be available in advance. This method is
implemented on Gaussian and non-Gaussian multidimensional data, a number of UCI
datasets and Indian Pine dataset. Results are compared with NWFE in terms of
classification accuracy and execution time. For nearest neighbour classifier it
shows that this technique converges to NWFE at the end of learning process. In
addition, the computational complexity is reduced in comparison with NWFE in
terms of execution time.
| [
{
"version": "v1",
"created": "Wed, 26 Oct 2016 01:02:01 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Moghaddam",
"Hamid Abrishami",
""
],
[
"Raisi",
"Elaheh",
""
]
] | TITLE: Incremental Nonparametric Weighted Feature Extraction for OnlineSubspace
Pattern Classification
ABSTRACT: In this paper, a new online method based on nonparametric weighted feature
extraction (NWFE) is proposed. NWFE was introduced to enjoy optimum
characteristics of linear discriminant analysis (LDA) and nonparametric
discriminant analysis (NDA) while rectifying their drawbacks. It emphasizes the
points near decision boundary by putting greater weights on them and
deemphasizes other points. Incremental nonparametric weighted feature
extraction (INWFE) is the online version of NWFE. INWFE has advantages of NWFE
method such as extracting more than L-1 features in contrast to LDA. It is
independent of the class distribution and performs well in complex distributed
data. The effects of outliers are reduced due to the nature of its
nonparametric scatter matrix. Furthermore, it is possible to add new samples
asynchronously, i.e. whenever a new sample becomes available at any given time,
it can be added to the algorithm. This is useful for many real world
applications since all data cannot be available in advance. This method is
implemented on Gaussian and non-Gaussian multidimensional data, a number of UCI
datasets and Indian Pine dataset. Results are compared with NWFE in terms of
classification accuracy and execution time. For nearest neighbour classifier it
shows that this technique converges to NWFE at the end of learning process. In
addition, the computational complexity is reduced in comparison with NWFE in
terms of execution time.
| no_new_dataset | 0.943034 |
1610.08462 | Qian Chen | Qian Chen, Xiaodan Zhu, Zhenhua Ling, Si Wei, Hui Jiang | Distraction-Based Neural Networks for Document Summarization | Published in IJCAI-2016: the 25th International Joint Conference on
Artificial Intelligence | IJCAI, 2016 | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distributed representation learned with neural networks has recently shown to
be effective in modeling natural languages at fine granularities such as words,
phrases, and even sentences. Whether and how such an approach can be extended
to help model larger spans of text, e.g., documents, is intriguing, and further
investigation would still be desirable. This paper aims to enhance neural
network models for such a purpose. A typical problem of document-level modeling
is automatic summarization, which aims to model documents in order to generate
summaries. In this paper, we propose neural models to train computers not just
to pay attention to specific regions and content of input documents with
attention models, but also distract them to traverse between different content
of a document so as to better grasp the overall meaning for summarization.
Without engineering any features, we train the models on two large datasets.
The models achieve the state-of-the-art performance, and they significantly
benefit from the distraction modeling, particularly when input documents are
long.
| [
{
"version": "v1",
"created": "Wed, 26 Oct 2016 18:57:00 GMT"
}
] | 2016-10-27T00:00:00 | [
[
"Chen",
"Qian",
""
],
[
"Zhu",
"Xiaodan",
""
],
[
"Ling",
"Zhenhua",
""
],
[
"Wei",
"Si",
""
],
[
"Jiang",
"Hui",
""
]
] | TITLE: Distraction-Based Neural Networks for Document Summarization
ABSTRACT: Distributed representation learned with neural networks has recently shown to
be effective in modeling natural languages at fine granularities such as words,
phrases, and even sentences. Whether and how such an approach can be extended
to help model larger spans of text, e.g., documents, is intriguing, and further
investigation would still be desirable. This paper aims to enhance neural
network models for such a purpose. A typical problem of document-level modeling
is automatic summarization, which aims to model documents in order to generate
summaries. In this paper, we propose neural models to train computers not just
to pay attention to specific regions and content of input documents with
attention models, but also distract them to traverse between different content
of a document so as to better grasp the overall meaning for summarization.
Without engineering any features, we train the models on two large datasets.
The models achieve the state-of-the-art performance, and they significantly
benefit from the distraction modeling, particularly when input documents are
long.
| no_new_dataset | 0.943919 |
1511.04664 | Mohammad Abu Alsheikh | Mohammad Abu Alsheikh, Ahmed Selim, Dusit Niyato, Linda Doyle, Shaowei
Lin, Hwee-Pink Tan | Deep Activity Recognition Models with Triaxial Accelerometers | null | null | null | null | cs.LG cs.HC cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the widespread installation of accelerometers in almost all mobile
phones and wearable devices, activity recognition using accelerometers is still
immature due to the poor recognition accuracy of existing recognition methods
and the scarcity of labeled training data. We consider the problem of human
activity recognition using triaxial accelerometers and deep learning paradigms.
This paper shows that deep activity recognition models (a) provide better
recognition accuracy of human activities, (b) avoid the expensive design of
handcrafted features in existing systems, and (c) utilize the massive unlabeled
acceleration samples for unsupervised feature extraction. Moreover, a hybrid
approach of deep learning and hidden Markov models (DL-HMM) is presented for
sequential activity recognition. This hybrid approach integrates the
hierarchical representations of deep activity recognition models with the
stochastic modeling of temporal sequences in the hidden Markov models. We show
substantial recognition improvement on real world datasets over
state-of-the-art methods of human activity recognition using triaxial
accelerometers.
| [
{
"version": "v1",
"created": "Sun, 15 Nov 2015 06:23:40 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Oct 2016 07:39:29 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Alsheikh",
"Mohammad Abu",
""
],
[
"Selim",
"Ahmed",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Doyle",
"Linda",
""
],
[
"Lin",
"Shaowei",
""
],
[
"Tan",
"Hwee-Pink",
""
]
] | TITLE: Deep Activity Recognition Models with Triaxial Accelerometers
ABSTRACT: Despite the widespread installation of accelerometers in almost all mobile
phones and wearable devices, activity recognition using accelerometers is still
immature due to the poor recognition accuracy of existing recognition methods
and the scarcity of labeled training data. We consider the problem of human
activity recognition using triaxial accelerometers and deep learning paradigms.
This paper shows that deep activity recognition models (a) provide better
recognition accuracy of human activities, (b) avoid the expensive design of
handcrafted features in existing systems, and (c) utilize the massive unlabeled
acceleration samples for unsupervised feature extraction. Moreover, a hybrid
approach of deep learning and hidden Markov models (DL-HMM) is presented for
sequential activity recognition. This hybrid approach integrates the
hierarchical representations of deep activity recognition models with the
stochastic modeling of temporal sequences in the hidden Markov models. We show
substantial recognition improvement on real world datasets over
state-of-the-art methods of human activity recognition using triaxial
accelerometers.
| no_new_dataset | 0.948489 |
1511.08250 | Bernardino Romera-Paredes | Bernardino Romera-Paredes, Philip H. S. Torr | Recurrent Instance Segmentation | 14 pages (main paper). 24 pages including references and appendix | ECCV 2016. 14th European Conference on Computer Vision | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Instance segmentation is the problem of detecting and delineating each
distinct object of interest appearing in an image. Current instance
segmentation approaches consist of ensembles of modules that are trained
independently of each other, thus missing opportunities for joint learning.
Here we propose a new instance segmentation paradigm consisting in an
end-to-end method that learns how to segment instances sequentially. The model
is based on a recurrent neural network that sequentially finds objects and
their segmentations one at a time. This net is provided with a spatial memory
that keeps track of what pixels have been explained and allows occlusion
handling. In order to train the model we designed a principled loss function
that accurately represents the properties of the instance segmentation problem.
In the experiments carried out, we found that our method outperforms recent
approaches on multiple person segmentation, and all state of the art approaches
on the Plant Phenotyping dataset for leaf counting.
| [
{
"version": "v1",
"created": "Wed, 25 Nov 2015 23:28:14 GMT"
},
{
"version": "v2",
"created": "Tue, 5 Apr 2016 22:45:04 GMT"
},
{
"version": "v3",
"created": "Mon, 24 Oct 2016 23:57:19 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Romera-Paredes",
"Bernardino",
""
],
[
"Torr",
"Philip H. S.",
""
]
] | TITLE: Recurrent Instance Segmentation
ABSTRACT: Instance segmentation is the problem of detecting and delineating each
distinct object of interest appearing in an image. Current instance
segmentation approaches consist of ensembles of modules that are trained
independently of each other, thus missing opportunities for joint learning.
Here we propose a new instance segmentation paradigm consisting in an
end-to-end method that learns how to segment instances sequentially. The model
is based on a recurrent neural network that sequentially finds objects and
their segmentations one at a time. This net is provided with a spatial memory
that keeps track of what pixels have been explained and allows occlusion
handling. In order to train the model we designed a principled loss function
that accurately represents the properties of the instance segmentation problem.
In the experiments carried out, we found that our method outperforms recent
approaches on multiple person segmentation, and all state of the art approaches
on the Plant Phenotyping dataset for leaf counting.
| no_new_dataset | 0.950365 |
1604.06045 | Jason Weston | Jason Weston | Dialog-based Language Learning | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A long-term goal of machine learning research is to build an intelligent
dialog agent. Most research in natural language understanding has focused on
learning from fixed training sets of labeled data, with supervision either at
the word level (tagging, parsing tasks) or sentence level (question answering,
machine translation). This kind of supervision is not realistic of how humans
learn, where language is both learned by, and used for, communication. In this
work, we study dialog-based language learning, where supervision is given
naturally and implicitly in the response of the dialog partner during the
conversation. We study this setup in two domains: the bAbI dataset of (Weston
et al., 2015) and large-scale question answering from (Dodge et al., 2015). We
evaluate a set of baseline learning strategies on these tasks, and show that a
novel model incorporating predictive lookahead is a promising approach for
learning from a teacher's response. In particular, a surprising result is that
it can learn to answer questions correctly without any reward-based supervision
at all.
| [
{
"version": "v1",
"created": "Wed, 20 Apr 2016 18:06:49 GMT"
},
{
"version": "v2",
"created": "Mon, 25 Apr 2016 18:27:03 GMT"
},
{
"version": "v3",
"created": "Wed, 18 May 2016 14:02:08 GMT"
},
{
"version": "v4",
"created": "Fri, 20 May 2016 02:53:30 GMT"
},
{
"version": "v5",
"created": "Tue, 23 Aug 2016 18:46:16 GMT"
},
{
"version": "v6",
"created": "Wed, 28 Sep 2016 21:30:27 GMT"
},
{
"version": "v7",
"created": "Mon, 24 Oct 2016 20:00:13 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Weston",
"Jason",
""
]
] | TITLE: Dialog-based Language Learning
ABSTRACT: A long-term goal of machine learning research is to build an intelligent
dialog agent. Most research in natural language understanding has focused on
learning from fixed training sets of labeled data, with supervision either at
the word level (tagging, parsing tasks) or sentence level (question answering,
machine translation). This kind of supervision is not realistic of how humans
learn, where language is both learned by, and used for, communication. In this
work, we study dialog-based language learning, where supervision is given
naturally and implicitly in the response of the dialog partner during the
conversation. We study this setup in two domains: the bAbI dataset of (Weston
et al., 2015) and large-scale question answering from (Dodge et al., 2015). We
evaluate a set of baseline learning strategies on these tasks, and show that a
novel model incorporating predictive lookahead is a promising approach for
learning from a teacher's response. In particular, a surprising result is that
it can learn to answer questions correctly without any reward-based supervision
at all.
| no_new_dataset | 0.947235 |
1605.06240 | Yangyan Li | Yangyan Li and Soeren Pirk and Hao Su and Charles R. Qi and Leonidas
J. Guibas | FPNN: Field Probing Neural Networks for 3D Data | To appear in NIPS 2016 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Building discriminative representations for 3D data has been an important
task in computer graphics and computer vision research. Convolutional Neural
Networks (CNNs) have shown to operate on 2D images with great success for a
variety of tasks. Lifting convolution operators to 3D (3DCNNs) seems like a
plausible and promising next step. Unfortunately, the computational complexity
of 3D CNNs grows cubically with respect to voxel resolution. Moreover, since
most 3D geometry representations are boundary based, occupied regions do not
increase proportionately with the size of the discretization, resulting in
wasted computation. In this work, we represent 3D spaces as volumetric fields,
and propose a novel design that employs field probing filters to efficiently
extract features from them. Each field probing filter is a set of probing
points --- sensors that perceive the space. Our learning algorithm optimizes
not only the weights associated with the probing points, but also their
locations, which deforms the shape of the probing filters and adaptively
distributes them in 3D space. The optimized probing points sense the 3D space
"intelligently", rather than operating blindly over the entire domain. We show
that field probing is significantly more efficient than 3DCNNs, while providing
state-of-the-art performance, on classification tasks for 3D object recognition
benchmark datasets.
| [
{
"version": "v1",
"created": "Fri, 20 May 2016 08:15:57 GMT"
},
{
"version": "v2",
"created": "Mon, 29 Aug 2016 07:34:49 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Oct 2016 03:59:16 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Li",
"Yangyan",
""
],
[
"Pirk",
"Soeren",
""
],
[
"Su",
"Hao",
""
],
[
"Qi",
"Charles R.",
""
],
[
"Guibas",
"Leonidas J.",
""
]
] | TITLE: FPNN: Field Probing Neural Networks for 3D Data
ABSTRACT: Building discriminative representations for 3D data has been an important
task in computer graphics and computer vision research. Convolutional Neural
Networks (CNNs) have shown to operate on 2D images with great success for a
variety of tasks. Lifting convolution operators to 3D (3DCNNs) seems like a
plausible and promising next step. Unfortunately, the computational complexity
of 3D CNNs grows cubically with respect to voxel resolution. Moreover, since
most 3D geometry representations are boundary based, occupied regions do not
increase proportionately with the size of the discretization, resulting in
wasted computation. In this work, we represent 3D spaces as volumetric fields,
and propose a novel design that employs field probing filters to efficiently
extract features from them. Each field probing filter is a set of probing
points --- sensors that perceive the space. Our learning algorithm optimizes
not only the weights associated with the probing points, but also their
locations, which deforms the shape of the probing filters and adaptively
distributes them in 3D space. The optimized probing points sense the 3D space
"intelligently", rather than operating blindly over the entire domain. We show
that field probing is significantly more efficient than 3DCNNs, while providing
state-of-the-art performance, on classification tasks for 3D object recognition
benchmark datasets.
| no_new_dataset | 0.95222 |
1605.06265 | Julien Mairal | Julien Mairal | End-to-End Kernel Learning with Supervised Convolutional Kernel Networks | to appear in Advances in Neural Information Processing Systems (NIPS) | null | null | null | stat.ML cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce a new image representation based on a multilayer
kernel machine. Unlike traditional kernel methods where data representation is
decoupled from the prediction task, we learn how to shape the kernel with
supervision. We proceed by first proposing improvements of the
recently-introduced convolutional kernel networks (CKNs) in the context of
unsupervised learning; then, we derive backpropagation rules to take advantage
of labeled training data. The resulting model is a new type of convolutional
neural network, where optimizing the filters at each layer is equivalent to
learning a linear subspace in a reproducing kernel Hilbert space (RKHS). We
show that our method achieves reasonably competitive performance for image
classification on some standard "deep learning" datasets such as CIFAR-10 and
SVHN, and also for image super-resolution, demonstrating the applicability of
our approach to a large variety of image-related tasks.
| [
{
"version": "v1",
"created": "Fri, 20 May 2016 09:52:14 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Oct 2016 12:52:50 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Mairal",
"Julien",
""
]
] | TITLE: End-to-End Kernel Learning with Supervised Convolutional Kernel Networks
ABSTRACT: In this paper, we introduce a new image representation based on a multilayer
kernel machine. Unlike traditional kernel methods where data representation is
decoupled from the prediction task, we learn how to shape the kernel with
supervision. We proceed by first proposing improvements of the
recently-introduced convolutional kernel networks (CKNs) in the context of
unsupervised learning; then, we derive backpropagation rules to take advantage
of labeled training data. The resulting model is a new type of convolutional
neural network, where optimizing the filters at each layer is equivalent to
learning a linear subspace in a reproducing kernel Hilbert space (RKHS). We
show that our method achieves reasonably competitive performance for image
classification on some standard "deep learning" datasets such as CIFAR-10 and
SVHN, and also for image super-resolution, demonstrating the applicability of
our approach to a large variety of image-related tasks.
| no_new_dataset | 0.949809 |
1610.07677 | Edward Yu | Edward Yu, Parth Parekh | A Bayesian Ensemble for Unsupervised Anomaly Detection | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Methods for unsupervised anomaly detection suffer from the fact that the data
is unlabeled, making it difficult to assess the optimality of detection
algorithms. Ensemble learning has shown exceptional results in classification
and clustering problems, but has not seen as much research in the context of
outlier detection. Existing methods focus on combining output scores of
individual detectors, but this leads to outputs that are not easily
interpretable. In this paper, we introduce a theoretical foundation for
combining individual detectors with Bayesian classifier combination. Not only
are posterior distributions easily interpreted as the probability distribution
of anomalies, but bias, variance, and individual error rates of detectors are
all easily obtained. Performance on real-world datasets shows high accuracy
across varied types of time series data.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 23:07:16 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Yu",
"Edward",
""
],
[
"Parekh",
"Parth",
""
]
] | TITLE: A Bayesian Ensemble for Unsupervised Anomaly Detection
ABSTRACT: Methods for unsupervised anomaly detection suffer from the fact that the data
is unlabeled, making it difficult to assess the optimality of detection
algorithms. Ensemble learning has shown exceptional results in classification
and clustering problems, but has not seen as much research in the context of
outlier detection. Existing methods focus on combining output scores of
individual detectors, but this leads to outputs that are not easily
interpretable. In this paper, we introduce a theoretical foundation for
combining individual detectors with Bayesian classifier combination. Not only
are posterior distributions easily interpreted as the probability distribution
of anomalies, but bias, variance, and individual error rates of detectors are
all easily obtained. Performance on real-world datasets shows high accuracy
across varied types of time series data.
| no_new_dataset | 0.949623 |
1610.07722 | Ioakeim Perros | Ioakeim Perros and Robert Chen and Richard Vuduc and Jimeng Sun | Sparse Hierarchical Tucker Factorization and its Application to
Healthcare | This is an extended version of a paper presented at the 15th IEEE
International Conference on Data Mining (ICDM 2015) | null | null | null | cs.LG cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new tensor factorization method, called the Sparse
Hierarchical-Tucker (Sparse H-Tucker), for sparse and high-order data tensors.
Sparse H-Tucker is inspired by its namesake, the classical Hierarchical Tucker
method, which aims to compute a tree-structured factorization of an input data
set that may be readily interpreted by a domain expert. However, Sparse
H-Tucker uses a nested sampling technique to overcome a key scalability problem
in Hierarchical Tucker, which is the creation of an unwieldy intermediate dense
core tensor; the result of our approach is a faster, more space-efficient, and
more accurate method. We extensively test our method on a real healthcare
dataset, which is collected from 30K patients and results in an 18th order
sparse data tensor. Unlike competing methods, Sparse H-Tucker can analyze the
full data set on a single multi-threaded machine. It can also do so more
accurately and in less time than the state-of-the-art: on a 12th order subset
of the input data, Sparse H-Tucker is 18x more accurate and 7.5x faster than a
previously state-of-the-art method. Even for analyzing low order tensors (e.g.,
4-order), our method requires close to an order of magnitude less time and over
two orders of magnitude less memory, as compared to traditional tensor
factorization methods such as CP and Tucker. Moreover, we observe that Sparse
H-Tucker scales nearly linearly in the number of non-zero tensor elements. The
resulting model also provides an interpretable disease hierarchy, which is
confirmed by a clinical expert.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 04:08:11 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Perros",
"Ioakeim",
""
],
[
"Chen",
"Robert",
""
],
[
"Vuduc",
"Richard",
""
],
[
"Sun",
"Jimeng",
""
]
] | TITLE: Sparse Hierarchical Tucker Factorization and its Application to
Healthcare
ABSTRACT: We propose a new tensor factorization method, called the Sparse
Hierarchical-Tucker (Sparse H-Tucker), for sparse and high-order data tensors.
Sparse H-Tucker is inspired by its namesake, the classical Hierarchical Tucker
method, which aims to compute a tree-structured factorization of an input data
set that may be readily interpreted by a domain expert. However, Sparse
H-Tucker uses a nested sampling technique to overcome a key scalability problem
in Hierarchical Tucker, which is the creation of an unwieldy intermediate dense
core tensor; the result of our approach is a faster, more space-efficient, and
more accurate method. We extensively test our method on a real healthcare
dataset, which is collected from 30K patients and results in an 18th order
sparse data tensor. Unlike competing methods, Sparse H-Tucker can analyze the
full data set on a single multi-threaded machine. It can also do so more
accurately and in less time than the state-of-the-art: on a 12th order subset
of the input data, Sparse H-Tucker is 18x more accurate and 7.5x faster than a
previously state-of-the-art method. Even for analyzing low order tensors (e.g.,
4-order), our method requires close to an order of magnitude less time and over
two orders of magnitude less memory, as compared to traditional tensor
factorization methods such as CP and Tucker. Moreover, we observe that Sparse
H-Tucker scales nearly linearly in the number of non-zero tensor elements. The
resulting model also provides an interpretable disease hierarchy, which is
confirmed by a clinical expert.
| no_new_dataset | 0.949995 |
1610.07732 | Anja Gruenheid | Anja Gruenheid, Donald Kossmann, Divesh Srivastava | Online Event Integration with StoryPivot | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern data integration systems need to process large amounts of data from a
variety of data sources and with real-time integration constraints. They are
not only employed in enterprises for managing internal data but are also used
for a variety of web services that use techniques such as entity resolution or
data cleaning in live systems. In this work, we discuss a new generation of
data integration systems that operate on (un-)structured data in an online
setting, i.e., systems which process continuously modified datasets upon which
the integration task is based. We use as an example of such a system an online
event integration system called StoryPivot. It observes events extracted from
news articles in data sources such as the 'Guardian' or the 'Washington Post'
which are integrated to show users the evolution of real-world stories over
time. The design decisions for StoryPivot are influenced by the trade-off
between maintaining high quality integration results while at the same time
building a system that processes and integrates events in near real-time. We
evaluate our design decisions with experiments on two real-world datasets and
generalize our findings to other data integration tasks that have a similar
system setup.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 05:10:18 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Gruenheid",
"Anja",
""
],
[
"Kossmann",
"Donald",
""
],
[
"Srivastava",
"Divesh",
""
]
] | TITLE: Online Event Integration with StoryPivot
ABSTRACT: Modern data integration systems need to process large amounts of data from a
variety of data sources and with real-time integration constraints. They are
not only employed in enterprises for managing internal data but are also used
for a variety of web services that use techniques such as entity resolution or
data cleaning in live systems. In this work, we discuss a new generation of
data integration systems that operate on (un-)structured data in an online
setting, i.e., systems which process continuously modified datasets upon which
the integration task is based. We use as an example of such a system an online
event integration system called StoryPivot. It observes events extracted from
news articles in data sources such as the 'Guardian' or the 'Washington Post'
which are integrated to show users the evolution of real-world stories over
time. The design decisions for StoryPivot are influenced by the trade-off
between maintaining high quality integration results while at the same time
building a system that processes and integrates events in near real-time. We
evaluate our design decisions with experiments on two real-world datasets and
generalize our findings to other data integration tasks that have a similar
system setup.
| no_new_dataset | 0.94625 |
1610.07752 | Mrutyunjaya Panda | Mrutyunjaya Panda | Big Models for Big Data using Multi objective averaged one dependence
estimators | 21 pages, 2 Figures, 10 tables | null | null | null | cs.NE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Even though, many researchers tried to explore the various possibilities on
multi objective feature selection, still it is yet to be explored with best of
its capabilities in data mining applications rather than going for developing
new ones. In this paper, multi-objective evolutionary algorithm ENORA is used
to select the features in a multi-class classification problem. The fusion of
AnDE (averaged n-dependence estimators) with n=1, a variant of naive Bayes with
efficient feature selection by ENORA is performed in order to obtain a fast
hybrid classifier which can effectively learn from big data. This method aims
at solving the problem of finding optimal feature subset from full data which
at present still remains to be a difficult problem. The efficacy of the
obtained classifier is extensively evaluated with a range of most popular 21
real world dataset, ranging from small to big. The results obtained are
encouraging in terms of time, Root mean square error, zero-one loss and
classification accuracy.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 07:11:11 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Panda",
"Mrutyunjaya",
""
]
] | TITLE: Big Models for Big Data using Multi objective averaged one dependence
estimators
ABSTRACT: Even though, many researchers tried to explore the various possibilities on
multi objective feature selection, still it is yet to be explored with best of
its capabilities in data mining applications rather than going for developing
new ones. In this paper, multi-objective evolutionary algorithm ENORA is used
to select the features in a multi-class classification problem. The fusion of
AnDE (averaged n-dependence estimators) with n=1, a variant of naive Bayes with
efficient feature selection by ENORA is performed in order to obtain a fast
hybrid classifier which can effectively learn from big data. This method aims
at solving the problem of finding optimal feature subset from full data which
at present still remains to be a difficult problem. The efficacy of the
obtained classifier is extensively evaluated with a range of most popular 21
real world dataset, ranging from small to big. The results obtained are
encouraging in terms of time, Root mean square error, zero-one loss and
classification accuracy.
| no_new_dataset | 0.947817 |
1610.07758 | Malay Bhattacharyya | Abhisek Dash, Sujoy Chatterjee, Tripti Prasad, and Malay Bhattacharyya | Image Clustering without Ground Truth | GroupSight Workshop, Fourth AAAI Conference on Human Computation and
Crowdsourcing (HCOMP 2016), Austin, USA | null | null | null | cs.HC cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cluster analysis has become one of the most exercised research areas over the
past few decades in computer science. As a consequence, numerous clustering
algorithms have already been developed to find appropriate partitions of a set
of objects. Given multiple such clustering solutions, it is a challenging task
to obtain an ensemble of these solutions. This becomes more challenging when
the ground truth about the number of clusters is unavailable. In this paper, we
introduce a crowd-powered model to collect solutions of image clustering from
the general crowd and pose it as a clustering ensemble problem with variable
number of clusters. The varying number of clusters basically reflects the crowd
workers' perspective toward a particular set of objects. We allow a set of
crowd workers to independently cluster the images as per their perceptions. We
address the problem by finding out centroid of the clusters using an
appropriate distance measure and prioritize the likelihood of similarity of the
individual cluster sets. The effectiveness of the proposed method is
demonstrated by applying it on multiple artificial datasets obtained from
crowd.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 07:34:47 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Dash",
"Abhisek",
""
],
[
"Chatterjee",
"Sujoy",
""
],
[
"Prasad",
"Tripti",
""
],
[
"Bhattacharyya",
"Malay",
""
]
] | TITLE: Image Clustering without Ground Truth
ABSTRACT: Cluster analysis has become one of the most exercised research areas over the
past few decades in computer science. As a consequence, numerous clustering
algorithms have already been developed to find appropriate partitions of a set
of objects. Given multiple such clustering solutions, it is a challenging task
to obtain an ensemble of these solutions. This becomes more challenging when
the ground truth about the number of clusters is unavailable. In this paper, we
introduce a crowd-powered model to collect solutions of image clustering from
the general crowd and pose it as a clustering ensemble problem with variable
number of clusters. The varying number of clusters basically reflects the crowd
workers' perspective toward a particular set of objects. We allow a set of
crowd workers to independently cluster the images as per their perceptions. We
address the problem by finding out centroid of the clusters using an
appropriate distance measure and prioritize the likelihood of similarity of the
individual cluster sets. The effectiveness of the proposed method is
demonstrated by applying it on multiple artificial datasets obtained from
crowd.
| no_new_dataset | 0.949153 |
1610.07809 | Florian Boudin | Florian Boudin, Hugo Mougard, Damien Cram | How Document Pre-processing affects Keyphrase Extraction Performance | Accepted at the COLING 2016 Workshop on Noisy User-generated Text
(WNUT) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The SemEval-2010 benchmark dataset has brought renewed attention to the task
of automatic keyphrase extraction. This dataset is made up of scientific
articles that were automatically converted from PDF format to plain text and
thus require careful preprocessing so that irrevelant spans of text do not
negatively affect keyphrase extraction performance. In previous work, a wide
range of document preprocessing techniques were described but their impact on
the overall performance of keyphrase extraction models is still unexplored.
Here, we re-assess the performance of several keyphrase extraction models and
measure their robustness against increasingly sophisticated levels of document
preprocessing.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 09:59:13 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Boudin",
"Florian",
""
],
[
"Mougard",
"Hugo",
""
],
[
"Cram",
"Damien",
""
]
] | TITLE: How Document Pre-processing affects Keyphrase Extraction Performance
ABSTRACT: The SemEval-2010 benchmark dataset has brought renewed attention to the task
of automatic keyphrase extraction. This dataset is made up of scientific
articles that were automatically converted from PDF format to plain text and
thus require careful preprocessing so that irrevelant spans of text do not
negatively affect keyphrase extraction performance. In previous work, a wide
range of document preprocessing techniques were described but their impact on
the overall performance of keyphrase extraction models is still unexplored.
Here, we re-assess the performance of several keyphrase extraction models and
measure their robustness against increasingly sophisticated levels of document
preprocessing.
| no_new_dataset | 0.919859 |
1610.07882 | Michael Blot | Michael Blot, Matthieu Cord, Nicolas Thome | Maxmin convolutional neural networks for image classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Convolutional neural networks (CNN) are widely used in computer vision,
especially in image classification. However, the way in which information and
invariance properties are encoded through in deep CNN architectures is still an
open question. In this paper, we propose to modify the standard convo- lutional
block of CNN in order to transfer more information layer after layer while
keeping some invariance within the net- work. Our main idea is to exploit both
positive and negative high scores obtained in the convolution maps. This behav-
ior is obtained by modifying the traditional activation func- tion step before
pooling. We are doubling the maps with spe- cific activations functions, called
MaxMin strategy, in order to achieve our pipeline. Extensive experiments on two
classical datasets, MNIST and CIFAR-10, show that our deep MaxMin convolutional
net outperforms standard CNN.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 14:04:11 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Blot",
"Michael",
""
],
[
"Cord",
"Matthieu",
""
],
[
"Thome",
"Nicolas",
""
]
] | TITLE: Maxmin convolutional neural networks for image classification
ABSTRACT: Convolutional neural networks (CNN) are widely used in computer vision,
especially in image classification. However, the way in which information and
invariance properties are encoded through in deep CNN architectures is still an
open question. In this paper, we propose to modify the standard convo- lutional
block of CNN in order to transfer more information layer after layer while
keeping some invariance within the net- work. Our main idea is to exploit both
positive and negative high scores obtained in the convolution maps. This behav-
ior is obtained by modifying the traditional activation func- tion step before
pooling. We are doubling the maps with spe- cific activations functions, called
MaxMin strategy, in order to achieve our pipeline. Extensive experiments on two
classical datasets, MNIST and CIFAR-10, show that our deep MaxMin convolutional
net outperforms standard CNN.
| no_new_dataset | 0.951997 |
1610.07918 | Yossi Adi | Yossi Adi, Joseph Keshet, Emily Cibelli, Matthew Goldrick | Sequence Segmentation Using Joint RNN and Structured Prediction Models | under review | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe and analyze a simple and effective algorithm for sequence
segmentation applied to speech processing tasks. We propose a neural
architecture that is composed of two modules trained jointly: a recurrent
neural network (RNN) module and a structured prediction model. The RNN outputs
are considered as feature functions to the structured model. The overall model
is trained with a structured loss function which can be designed to the given
segmentation task. We demonstrate the effectiveness of our method by applying
it to two simple tasks commonly used in phonetic studies: word segmentation and
voice onset time segmentation. Results sug- gest the proposed model is superior
to previous methods, ob- taining state-of-the-art results on the tested
datasets.
| [
{
"version": "v1",
"created": "Tue, 25 Oct 2016 15:21:25 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Adi",
"Yossi",
""
],
[
"Keshet",
"Joseph",
""
],
[
"Cibelli",
"Emily",
""
],
[
"Goldrick",
"Matthew",
""
]
] | TITLE: Sequence Segmentation Using Joint RNN and Structured Prediction Models
ABSTRACT: We describe and analyze a simple and effective algorithm for sequence
segmentation applied to speech processing tasks. We propose a neural
architecture that is composed of two modules trained jointly: a recurrent
neural network (RNN) module and a structured prediction model. The RNN outputs
are considered as feature functions to the structured model. The overall model
is trained with a structured loss function which can be designed to the given
segmentation task. We demonstrate the effectiveness of our method by applying
it to two simple tasks commonly used in phonetic studies: word segmentation and
voice onset time segmentation. Results sug- gest the proposed model is superior
to previous methods, ob- taining state-of-the-art results on the tested
datasets.
| no_new_dataset | 0.952706 |
1610.08015 | Nicola Wadeson Dr | Nicola Wadeson, Mark Basham | Savu: A Python-based, MPI Framework for Simultaneous Processing of
Multiple, N-dimensional, Large Tomography Datasets | 10 pages, 10 figures, 1 table | null | null | null | cs.DC cs.CV cs.DB | http://creativecommons.org/licenses/by/4.0/ | Diamond Light Source (DLS), the UK synchrotron facility, attracts scientists
from across the world to perform ground-breaking x-ray experiments. With over
3000 scientific users per year, vast amounts of data are collected across the
experimental beamlines, with the highest volume of data collected during
tomographic imaging experiments. A growing interest in tomography as an imaging
technique, has led to an expansion in the range of experiments performed, in
addition to a growth in the size of the data per experiment.
Savu is a portable, flexible, scientific processing pipeline capable of
processing multiple, n-dimensional datasets in serial on a PC, or in parallel
across a cluster. Developed at DLS, and successfully deployed across the
beamlines, it uses a modular plugin format to enable experiment-specific
processing and utilises parallel HDF5 to remove RAM restrictions. The Savu
design, described throughout this paper, focuses on easy integration of
existing and new functionality, flexibility and ease of use for users and
developers alike.
| [
{
"version": "v1",
"created": "Mon, 24 Oct 2016 13:22:09 GMT"
}
] | 2016-10-26T00:00:00 | [
[
"Wadeson",
"Nicola",
""
],
[
"Basham",
"Mark",
""
]
] | TITLE: Savu: A Python-based, MPI Framework for Simultaneous Processing of
Multiple, N-dimensional, Large Tomography Datasets
ABSTRACT: Diamond Light Source (DLS), the UK synchrotron facility, attracts scientists
from across the world to perform ground-breaking x-ray experiments. With over
3000 scientific users per year, vast amounts of data are collected across the
experimental beamlines, with the highest volume of data collected during
tomographic imaging experiments. A growing interest in tomography as an imaging
technique, has led to an expansion in the range of experiments performed, in
addition to a growth in the size of the data per experiment.
Savu is a portable, flexible, scientific processing pipeline capable of
processing multiple, n-dimensional datasets in serial on a PC, or in parallel
across a cluster. Developed at DLS, and successfully deployed across the
beamlines, it uses a modular plugin format to enable experiment-specific
processing and utilises parallel HDF5 to remove RAM restrictions. The Savu
design, described throughout this paper, focuses on easy integration of
existing and new functionality, flexibility and ease of use for users and
developers alike.
| no_new_dataset | 0.945851 |
1511.04404 | Oncel Tuzel | Oncel Tuzel and Tim K. Marks and Salil Tambe | Robust Face Alignment Using a Mixture of Invariant Experts | 17 pages, 6 figures | Proceedings of 14th European Conference on Computer Vision (ECCV),
Amsterdam, The Netherlands, October 11-14, 2016, pp 825-841 | 10.1007/978-3-319-46454-1_50 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face alignment, which is the task of finding the locations of a set of facial
landmark points in an image of a face, is useful in widespread application
areas. Face alignment is particularly challenging when there are large
variations in pose (in-plane and out-of-plane rotations) and facial expression.
To address this issue, we propose a cascade in which each stage consists of a
mixture of regression experts. Each expert learns a customized regression model
that is specialized to a different subset of the joint space of pose and
expressions. The system is invariant to a predefined class of transformations
(e.g., affine), because the input is transformed to match each expert's
prototype shape before the regression is applied. We also present a method to
include deformation constraints within the discriminative alignment framework,
which makes our algorithm more robust. Our algorithm significantly outperforms
previous methods on publicly available face alignment datasets.
| [
{
"version": "v1",
"created": "Fri, 13 Nov 2015 19:14:51 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Oct 2016 18:31:06 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Tuzel",
"Oncel",
""
],
[
"Marks",
"Tim K.",
""
],
[
"Tambe",
"Salil",
""
]
] | TITLE: Robust Face Alignment Using a Mixture of Invariant Experts
ABSTRACT: Face alignment, which is the task of finding the locations of a set of facial
landmark points in an image of a face, is useful in widespread application
areas. Face alignment is particularly challenging when there are large
variations in pose (in-plane and out-of-plane rotations) and facial expression.
To address this issue, we propose a cascade in which each stage consists of a
mixture of regression experts. Each expert learns a customized regression model
that is specialized to a different subset of the joint space of pose and
expressions. The system is invariant to a predefined class of transformations
(e.g., affine), because the input is transformed to match each expert's
prototype shape before the regression is applied. We also present a method to
include deformation constraints within the discriminative alignment framework,
which makes our algorithm more robust. Our algorithm significantly outperforms
previous methods on publicly available face alignment datasets.
| no_new_dataset | 0.952926 |
1511.05616 | Hexiang Hu | Hexiang Hu, Guang-Tong Zhou, Zhiwei Deng, Zicheng Liao, Greg Mori | Learning Structured Inference Neural Networks with Label Relations | Conference on Computer Vision and Pattern Recognition(CVPR) 2016 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Images of scenes have various objects as well as abundant attributes, and
diverse levels of visual categorization are possible. A natural image could be
assigned with fine-grained labels that describe major components,
coarse-grained labels that depict high level abstraction or a set of labels
that reveal attributes. Such categorization at different concept layers can be
modeled with label graphs encoding label information. In this paper, we exploit
this rich information with a state-of-art deep learning framework, and propose
a generic structured model that leverages diverse label relations to improve
image classification performance. Our approach employs a novel stacked label
prediction neural network, capturing both inter-level and intra-level label
semantics. We evaluate our method on benchmark image datasets, and empirical
results illustrate the efficacy of our model.
| [
{
"version": "v1",
"created": "Tue, 17 Nov 2015 23:22:25 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Nov 2015 06:13:16 GMT"
},
{
"version": "v3",
"created": "Fri, 8 Apr 2016 05:04:52 GMT"
},
{
"version": "v4",
"created": "Mon, 24 Oct 2016 18:20:20 GMT"
}
] | 2016-10-25T00:00:00 | [
[
"Hu",
"Hexiang",
""
],
[
"Zhou",
"Guang-Tong",
""
],
[
"Deng",
"Zhiwei",
""
],
[
"Liao",
"Zicheng",
""
],
[
"Mori",
"Greg",
""
]
] | TITLE: Learning Structured Inference Neural Networks with Label Relations
ABSTRACT: Images of scenes have various objects as well as abundant attributes, and
diverse levels of visual categorization are possible. A natural image could be
assigned with fine-grained labels that describe major components,
coarse-grained labels that depict high level abstraction or a set of labels
that reveal attributes. Such categorization at different concept layers can be
modeled with label graphs encoding label information. In this paper, we exploit
this rich information with a state-of-art deep learning framework, and propose
a generic structured model that leverages diverse label relations to improve
image classification performance. Our approach employs a novel stacked label
prediction neural network, capturing both inter-level and intra-level label
semantics. We evaluate our method on benchmark image datasets, and empirical
results illustrate the efficacy of our model.
| no_new_dataset | 0.948251 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.