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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1412.3040 | Albert Kim | Albert Kim, Eric Blais, Aditya Parameswaran, Piotr Indyk, Sam Madden,
Ronitt Rubinfeld | Rapid Sampling for Visualizations with Ordering Guarantees | Tech Report. 17 pages. Condensed version to appear in VLDB Vol. 8 No.
5 | null | null | null | cs.DB | http://creativecommons.org/licenses/by-nc-sa/3.0/ | Visualizations are frequently used as a means to understand trends and gather
insights from datasets, but often take a long time to generate. In this paper,
we focus on the problem of rapidly generating approximate visualizations while
preserving crucial visual proper- ties of interest to analysts. Our primary
focus will be on sampling algorithms that preserve the visual property of
ordering; our techniques will also apply to some other visual properties. For
instance, our algorithms can be used to generate an approximate visualization
of a bar chart very rapidly, where the comparisons between any two bars are
correct. We formally show that our sampling algorithms are generally applicable
and provably optimal in theory, in that they do not take more samples than
necessary to generate the visualizations with ordering guarantees. They also
work well in practice, correctly ordering output groups while taking orders of
magnitude fewer samples and much less time than conventional sampling schemes.
| [
{
"version": "v1",
"created": "Tue, 9 Dec 2014 18:08:26 GMT"
}
] | 2014-12-10T00:00:00 | [
[
"Kim",
"Albert",
""
],
[
"Blais",
"Eric",
""
],
[
"Parameswaran",
"Aditya",
""
],
[
"Indyk",
"Piotr",
""
],
[
"Madden",
"Sam",
""
],
[
"Rubinfeld",
"Ronitt",
""
]
] | TITLE: Rapid Sampling for Visualizations with Ordering Guarantees
ABSTRACT: Visualizations are frequently used as a means to understand trends and gather
insights from datasets, but often take a long time to generate. In this paper,
we focus on the problem of rapidly generating approximate visualizations while
preserving crucial visual proper- ties of interest to analysts. Our primary
focus will be on sampling algorithms that preserve the visual property of
ordering; our techniques will also apply to some other visual properties. For
instance, our algorithms can be used to generate an approximate visualization
of a bar chart very rapidly, where the comparisons between any two bars are
correct. We formally show that our sampling algorithms are generally applicable
and provably optimal in theory, in that they do not take more samples than
necessary to generate the visualizations with ordering guarantees. They also
work well in practice, correctly ordering output groups while taking orders of
magnitude fewer samples and much less time than conventional sampling schemes.
| no_new_dataset | 0.951594 |
1406.1485 | Li Yao | Tapani Raiko, Li Yao, Kyunghyun Cho and Yoshua Bengio | Iterative Neural Autoregressive Distribution Estimator (NADE-k) | Accepted at Neural Information Processing Systems (NIPS) 2014 | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training of the neural autoregressive density estimator (NADE) can be viewed
as doing one step of probabilistic inference on missing values in data. We
propose a new model that extends this inference scheme to multiple steps,
arguing that it is easier to learn to improve a reconstruction in $k$ steps
rather than to learn to reconstruct in a single inference step. The proposed
model is an unsupervised building block for deep learning that combines the
desirable properties of NADE and multi-predictive training: (1) Its test
likelihood can be computed analytically, (2) it is easy to generate independent
samples from it, and (3) it uses an inference engine that is a superset of
variational inference for Boltzmann machines. The proposed NADE-k is
competitive with the state-of-the-art in density estimation on the two datasets
tested.
| [
{
"version": "v1",
"created": "Thu, 5 Jun 2014 19:13:51 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Jun 2014 11:28:36 GMT"
},
{
"version": "v3",
"created": "Sat, 6 Dec 2014 00:22:00 GMT"
}
] | 2014-12-09T00:00:00 | [
[
"Raiko",
"Tapani",
""
],
[
"Yao",
"Li",
""
],
[
"Cho",
"Kyunghyun",
""
],
[
"Bengio",
"Yoshua",
""
]
] | TITLE: Iterative Neural Autoregressive Distribution Estimator (NADE-k)
ABSTRACT: Training of the neural autoregressive density estimator (NADE) can be viewed
as doing one step of probabilistic inference on missing values in data. We
propose a new model that extends this inference scheme to multiple steps,
arguing that it is easier to learn to improve a reconstruction in $k$ steps
rather than to learn to reconstruct in a single inference step. The proposed
model is an unsupervised building block for deep learning that combines the
desirable properties of NADE and multi-predictive training: (1) Its test
likelihood can be computed analytically, (2) it is easy to generate independent
samples from it, and (3) it uses an inference engine that is a superset of
variational inference for Boltzmann machines. The proposed NADE-k is
competitive with the state-of-the-art in density estimation on the two datasets
tested.
| no_new_dataset | 0.946892 |
1411.1792 | Jason Yosinski | Jason Yosinski, Jeff Clune, Yoshua Bengio, Hod Lipson | How transferable are features in deep neural networks? | To appear in Advances in Neural Information Processing Systems 27
(NIPS 2014) | Advances in Neural Information Processing Systems 27, pages
3320-3328. Dec. 2014 | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many deep neural networks trained on natural images exhibit a curious
phenomenon in common: on the first layer they learn features similar to Gabor
filters and color blobs. Such first-layer features appear not to be specific to
a particular dataset or task, but general in that they are applicable to many
datasets and tasks. Features must eventually transition from general to
specific by the last layer of the network, but this transition has not been
studied extensively. In this paper we experimentally quantify the generality
versus specificity of neurons in each layer of a deep convolutional neural
network and report a few surprising results. Transferability is negatively
affected by two distinct issues: (1) the specialization of higher layer neurons
to their original task at the expense of performance on the target task, which
was expected, and (2) optimization difficulties related to splitting networks
between co-adapted neurons, which was not expected. In an example network
trained on ImageNet, we demonstrate that either of these two issues may
dominate, depending on whether features are transferred from the bottom,
middle, or top of the network. We also document that the transferability of
features decreases as the distance between the base task and target task
increases, but that transferring features even from distant tasks can be better
than using random features. A final surprising result is that initializing a
network with transferred features from almost any number of layers can produce
a boost to generalization that lingers even after fine-tuning to the target
dataset.
| [
{
"version": "v1",
"created": "Thu, 6 Nov 2014 23:09:37 GMT"
}
] | 2014-12-09T00:00:00 | [
[
"Yosinski",
"Jason",
""
],
[
"Clune",
"Jeff",
""
],
[
"Bengio",
"Yoshua",
""
],
[
"Lipson",
"Hod",
""
]
] | TITLE: How transferable are features in deep neural networks?
ABSTRACT: Many deep neural networks trained on natural images exhibit a curious
phenomenon in common: on the first layer they learn features similar to Gabor
filters and color blobs. Such first-layer features appear not to be specific to
a particular dataset or task, but general in that they are applicable to many
datasets and tasks. Features must eventually transition from general to
specific by the last layer of the network, but this transition has not been
studied extensively. In this paper we experimentally quantify the generality
versus specificity of neurons in each layer of a deep convolutional neural
network and report a few surprising results. Transferability is negatively
affected by two distinct issues: (1) the specialization of higher layer neurons
to their original task at the expense of performance on the target task, which
was expected, and (2) optimization difficulties related to splitting networks
between co-adapted neurons, which was not expected. In an example network
trained on ImageNet, we demonstrate that either of these two issues may
dominate, depending on whether features are transferred from the bottom,
middle, or top of the network. We also document that the transferability of
features decreases as the distance between the base task and target task
increases, but that transferring features even from distant tasks can be better
than using random features. A final surprising result is that initializing a
network with transferred features from almost any number of layers can produce
a boost to generalization that lingers even after fine-tuning to the target
dataset.
| no_new_dataset | 0.942135 |
1412.2485 | Vikram Nathan | Vikram Nathan, Sharath Raghvendra | Accurate Streaming Support Vector Machines | 2 figures, 8 pages | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A widely-used tool for binary classification is the Support Vector Machine
(SVM), a supervised learning technique that finds the "maximum margin" linear
separator between the two classes. While SVMs have been well studied in the
batch (offline) setting, there is considerably less work on the streaming
(online) setting, which requires only a single pass over the data using
sub-linear space. Existing streaming algorithms are not yet competitive with
the batch implementation. In this paper, we use the formulation of the SVM as a
minimum enclosing ball (MEB) problem to provide a streaming SVM algorithm based
off of the blurred ball cover originally proposed by Agarwal and Sharathkumar.
Our implementation consistently outperforms existing streaming SVM approaches
and provides higher accuracies than libSVM on several datasets, thus making it
competitive with the standard SVM batch implementation.
| [
{
"version": "v1",
"created": "Mon, 8 Dec 2014 08:46:07 GMT"
}
] | 2014-12-09T00:00:00 | [
[
"Nathan",
"Vikram",
""
],
[
"Raghvendra",
"Sharath",
""
]
] | TITLE: Accurate Streaming Support Vector Machines
ABSTRACT: A widely-used tool for binary classification is the Support Vector Machine
(SVM), a supervised learning technique that finds the "maximum margin" linear
separator between the two classes. While SVMs have been well studied in the
batch (offline) setting, there is considerably less work on the streaming
(online) setting, which requires only a single pass over the data using
sub-linear space. Existing streaming algorithms are not yet competitive with
the batch implementation. In this paper, we use the formulation of the SVM as a
minimum enclosing ball (MEB) problem to provide a streaming SVM algorithm based
off of the blurred ball cover originally proposed by Agarwal and Sharathkumar.
Our implementation consistently outperforms existing streaming SVM approaches
and provides higher accuracies than libSVM on several datasets, thus making it
competitive with the standard SVM batch implementation.
| no_new_dataset | 0.948202 |
1412.2697 | Hocine Cherifi | Abdelkaher Ait Abdelouahad, Mohammed El Hassouni, Hocine Cherifi, and
Driss Aboutajdine | Image quality assessment measure based on natural image statistics in
the Tetrolet domain | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper deals with a reduced reference (RR) image quality measure based on
natural image statistics modeling. For this purpose, Tetrolet transform is used
since it provides a convenient way to capture local geometric structures. This
transform is applied to both reference and distorted images. Then, Gaussian
Scale Mixture (GSM) is proposed to model subbands in order to take account
statistical dependencies between tetrolet coefficients. In order to quantify
the visual degradation, a measure based on Kullback Leibler Divergence (KLD) is
provided. The proposed measure was tested on the Cornell VCL A-57 dataset and
compared with other measures according to FR-TV1 VQEG framework.
| [
{
"version": "v1",
"created": "Mon, 8 Dec 2014 18:48:26 GMT"
}
] | 2014-12-09T00:00:00 | [
[
"Abdelouahad",
"Abdelkaher Ait",
""
],
[
"Hassouni",
"Mohammed El",
""
],
[
"Cherifi",
"Hocine",
""
],
[
"Aboutajdine",
"Driss",
""
]
] | TITLE: Image quality assessment measure based on natural image statistics in
the Tetrolet domain
ABSTRACT: This paper deals with a reduced reference (RR) image quality measure based on
natural image statistics modeling. For this purpose, Tetrolet transform is used
since it provides a convenient way to capture local geometric structures. This
transform is applied to both reference and distorted images. Then, Gaussian
Scale Mixture (GSM) is proposed to model subbands in order to take account
statistical dependencies between tetrolet coefficients. In order to quantify
the visual degradation, a measure based on Kullback Leibler Divergence (KLD) is
provided. The proposed measure was tested on the Cornell VCL A-57 dataset and
compared with other measures according to FR-TV1 VQEG framework.
| no_new_dataset | 0.951863 |
1412.2716 | Paul Ginsparg | Daniel T. Citron and Paul Ginsparg | Patterns of Text Reuse in a Scientific Corpus | 6 pages, plus 10 pages of supplementary material. To appear in PNAS
(online 8 Dec 2014) | null | 10.1073/pnas.1415135111 | null | cs.DL physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the incidence of text "reuse" by researchers, via a systematic
pairwise comparison of the text content of all articles deposited to arXiv.org
from 1991--2012. We measure the global frequencies of three classes of text
reuse, and measure how chronic text reuse is distributed among authors in the
dataset. We infer a baseline for accepted practice, perhaps surprisingly
permissive compared with other societal contexts, and a clearly delineated set
of aberrant authors. We find a negative correlation between the amount of
reused text in an article and its influence, as measured by subsequent
citations. Finally, we consider the distribution of countries of origin of
articles containing large amounts of reused text.
| [
{
"version": "v1",
"created": "Mon, 8 Dec 2014 20:01:17 GMT"
}
] | 2014-12-09T00:00:00 | [
[
"Citron",
"Daniel T.",
""
],
[
"Ginsparg",
"Paul",
""
]
] | TITLE: Patterns of Text Reuse in a Scientific Corpus
ABSTRACT: We consider the incidence of text "reuse" by researchers, via a systematic
pairwise comparison of the text content of all articles deposited to arXiv.org
from 1991--2012. We measure the global frequencies of three classes of text
reuse, and measure how chronic text reuse is distributed among authors in the
dataset. We infer a baseline for accepted practice, perhaps surprisingly
permissive compared with other societal contexts, and a clearly delineated set
of aberrant authors. We find a negative correlation between the amount of
reused text in an article and its influence, as measured by subsequent
citations. Finally, we consider the distribution of countries of origin of
articles containing large amounts of reused text.
| no_new_dataset | 0.9462 |
1410.7454 | Neeraj Dhungel | Neeraj Dhungel, Gustavo Carneiro, Andrew P. Bradley | Deep Structured learning for mass segmentation from Mammograms | 4 pages, 2 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/3.0/ | In this paper, we present a novel method for the segmentation of breast
masses from mammograms exploring structured and deep learning. Specifically,
using structured support vector machine (SSVM), we formulate a model that
combines different types of potential functions, including one that classifies
image regions using deep learning. Our main goal with this work is to show the
accuracy and efficiency improvements that these relatively new techniques can
provide for the segmentation of breast masses from mammograms. We also propose
an easily reproducible quantitative analysis to as- sess the performance of
breast mass segmentation methodologies based on widely accepted accuracy and
running time measurements on public datasets, which will facilitate further
comparisons for this segmentation problem. In particular, we use two publicly
available datasets (DDSM-BCRP and INbreast) and propose the computa- tion of
the running time taken for the methodology to produce a mass segmentation given
an input image and the use of the Dice index to quantitatively measure the
segmentation accuracy. For both databases, we show that our proposed
methodology produces competitive results in terms of accuracy and running time.
| [
{
"version": "v1",
"created": "Mon, 27 Oct 2014 22:44:26 GMT"
},
{
"version": "v2",
"created": "Fri, 5 Dec 2014 01:21:39 GMT"
}
] | 2014-12-08T00:00:00 | [
[
"Dhungel",
"Neeraj",
""
],
[
"Carneiro",
"Gustavo",
""
],
[
"Bradley",
"Andrew P.",
""
]
] | TITLE: Deep Structured learning for mass segmentation from Mammograms
ABSTRACT: In this paper, we present a novel method for the segmentation of breast
masses from mammograms exploring structured and deep learning. Specifically,
using structured support vector machine (SSVM), we formulate a model that
combines different types of potential functions, including one that classifies
image regions using deep learning. Our main goal with this work is to show the
accuracy and efficiency improvements that these relatively new techniques can
provide for the segmentation of breast masses from mammograms. We also propose
an easily reproducible quantitative analysis to as- sess the performance of
breast mass segmentation methodologies based on widely accepted accuracy and
running time measurements on public datasets, which will facilitate further
comparisons for this segmentation problem. In particular, we use two publicly
available datasets (DDSM-BCRP and INbreast) and propose the computa- tion of
the running time taken for the methodology to produce a mass segmentation given
an input image and the use of the Dice index to quantitatively measure the
segmentation accuracy. For both databases, we show that our proposed
methodology produces competitive results in terms of accuracy and running time.
| no_new_dataset | 0.951594 |
1412.1842 | Max Jaderberg | Max Jaderberg, Karen Simonyan, Andrea Vedaldi, Andrew Zisserman | Reading Text in the Wild with Convolutional Neural Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we present an end-to-end system for text spotting -- localising
and recognising text in natural scene images -- and text based image retrieval.
This system is based on a region proposal mechanism for detection and deep
convolutional neural networks for recognition. Our pipeline uses a novel
combination of complementary proposal generation techniques to ensure high
recall, and a fast subsequent filtering stage for improving precision. For the
recognition and ranking of proposals, we train very large convolutional neural
networks to perform word recognition on the whole proposal region at the same
time, departing from the character classifier based systems of the past. These
networks are trained solely on data produced by a synthetic text generation
engine, requiring no human labelled data.
Analysing the stages of our pipeline, we show state-of-the-art performance
throughout. We perform rigorous experiments across a number of standard
end-to-end text spotting benchmarks and text-based image retrieval datasets,
showing a large improvement over all previous methods. Finally, we demonstrate
a real-world application of our text spotting system to allow thousands of
hours of news footage to be instantly searchable via a text query.
| [
{
"version": "v1",
"created": "Thu, 4 Dec 2014 21:14:59 GMT"
}
] | 2014-12-08T00:00:00 | [
[
"Jaderberg",
"Max",
""
],
[
"Simonyan",
"Karen",
""
],
[
"Vedaldi",
"Andrea",
""
],
[
"Zisserman",
"Andrew",
""
]
] | TITLE: Reading Text in the Wild with Convolutional Neural Networks
ABSTRACT: In this work we present an end-to-end system for text spotting -- localising
and recognising text in natural scene images -- and text based image retrieval.
This system is based on a region proposal mechanism for detection and deep
convolutional neural networks for recognition. Our pipeline uses a novel
combination of complementary proposal generation techniques to ensure high
recall, and a fast subsequent filtering stage for improving precision. For the
recognition and ranking of proposals, we train very large convolutional neural
networks to perform word recognition on the whole proposal region at the same
time, departing from the character classifier based systems of the past. These
networks are trained solely on data produced by a synthetic text generation
engine, requiring no human labelled data.
Analysing the stages of our pipeline, we show state-of-the-art performance
throughout. We perform rigorous experiments across a number of standard
end-to-end text spotting benchmarks and text-based image retrieval datasets,
showing a large improvement over all previous methods. Finally, we demonstrate
a real-world application of our text spotting system to allow thousands of
hours of news footage to be instantly searchable via a text query.
| no_new_dataset | 0.95018 |
1412.1888 | Rafi Muhammad | Muhammad Rafi, Farnaz Amin, Mohammad Shahid Shaikh | Document clustering using graph based document representation with
constraints | null | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Document clustering is an unsupervised approach in which a large collection
of documents (corpus) is subdivided into smaller, meaningful, identifiable, and
verifiable sub-groups (clusters). Meaningful representation of documents and
implicitly identifying the patterns, on which this separation is performed, is
the challenging part of document clustering. We have proposed a document
clustering technique using graph based document representation with
constraints. A graph data structure can easily capture the non-linear
relationships of nodes, document contains various feature terms that can be
non-linearly connected hence a graph can easily represents this information.
Constrains, are explicit conditions for document clustering where background
knowledge is use to set the direction for Linking or Not-Linking a set of
documents for a target clusters, thus guiding the clustering process. We deemed
clustering is an ill-define problem, there can be many clustering results.
Background knowledge can be used to drive the clustering algorithm in the right
direction. We have proposed three different types of constraints, Instance
level, corpus level and cluster level constraints. A new algorithm Constrained
HAC is also proposed which will incorporate Instance level constraints as prior
knowledge; it will guide the clustering process leading to better results.
Extensive set of experiments have been performed on both synthetic and standard
document clustering datasets, results are compared on standard clustering
measures like: purity, entropy and F-measure. Results clearly establish that
our proposed approach leads to improvement in cluster quality.
| [
{
"version": "v1",
"created": "Fri, 5 Dec 2014 03:47:34 GMT"
}
] | 2014-12-08T00:00:00 | [
[
"Rafi",
"Muhammad",
""
],
[
"Amin",
"Farnaz",
""
],
[
"Shaikh",
"Mohammad Shahid",
""
]
] | TITLE: Document clustering using graph based document representation with
constraints
ABSTRACT: Document clustering is an unsupervised approach in which a large collection
of documents (corpus) is subdivided into smaller, meaningful, identifiable, and
verifiable sub-groups (clusters). Meaningful representation of documents and
implicitly identifying the patterns, on which this separation is performed, is
the challenging part of document clustering. We have proposed a document
clustering technique using graph based document representation with
constraints. A graph data structure can easily capture the non-linear
relationships of nodes, document contains various feature terms that can be
non-linearly connected hence a graph can easily represents this information.
Constrains, are explicit conditions for document clustering where background
knowledge is use to set the direction for Linking or Not-Linking a set of
documents for a target clusters, thus guiding the clustering process. We deemed
clustering is an ill-define problem, there can be many clustering results.
Background knowledge can be used to drive the clustering algorithm in the right
direction. We have proposed three different types of constraints, Instance
level, corpus level and cluster level constraints. A new algorithm Constrained
HAC is also proposed which will incorporate Instance level constraints as prior
knowledge; it will guide the clustering process leading to better results.
Extensive set of experiments have been performed on both synthetic and standard
document clustering datasets, results are compared on standard clustering
measures like: purity, entropy and F-measure. Results clearly establish that
our proposed approach leads to improvement in cluster quality.
| no_new_dataset | 0.952442 |
1412.1908 | Rui Zhao | Rui Zhao, Wanli Ouyang, Xiaogang Wang | Person Re-identification by Saliency Learning | This manuscript has 14 pages with 25 figures, and a preliminary
version was published in ICCV 2013 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/3.0/ | Human eyes can recognize person identities based on small salient regions,
i.e. human saliency is distinctive and reliable in pedestrian matching across
disjoint camera views. However, such valuable information is often hidden when
computing similarities of pedestrian images with existing approaches. Inspired
by our user study result of human perception on human saliency, we propose a
novel perspective for person re-identification based on learning human saliency
and matching saliency distribution. The proposed saliency learning and matching
framework consists of four steps: (1) To handle misalignment caused by drastic
viewpoint change and pose variations, we apply adjacency constrained patch
matching to build dense correspondence between image pairs. (2) We propose two
alternative methods, i.e. K-Nearest Neighbors and One-class SVM, to estimate a
saliency score for each image patch, through which distinctive features stand
out without using identity labels in the training procedure. (3) saliency
matching is proposed based on patch matching. Matching patches with
inconsistent saliency brings penalty, and images of the same identity are
recognized by minimizing the saliency matching cost. (4) Furthermore, saliency
matching is tightly integrated with patch matching in a unified structural
RankSVM learning framework. The effectiveness of our approach is validated on
the VIPeR dataset and the CUHK01 dataset. Our approach outperforms the
state-of-the-art person re-identification methods on both datasets.
| [
{
"version": "v1",
"created": "Fri, 5 Dec 2014 07:33:48 GMT"
}
] | 2014-12-08T00:00:00 | [
[
"Zhao",
"Rui",
""
],
[
"Ouyang",
"Wanli",
""
],
[
"Wang",
"Xiaogang",
""
]
] | TITLE: Person Re-identification by Saliency Learning
ABSTRACT: Human eyes can recognize person identities based on small salient regions,
i.e. human saliency is distinctive and reliable in pedestrian matching across
disjoint camera views. However, such valuable information is often hidden when
computing similarities of pedestrian images with existing approaches. Inspired
by our user study result of human perception on human saliency, we propose a
novel perspective for person re-identification based on learning human saliency
and matching saliency distribution. The proposed saliency learning and matching
framework consists of four steps: (1) To handle misalignment caused by drastic
viewpoint change and pose variations, we apply adjacency constrained patch
matching to build dense correspondence between image pairs. (2) We propose two
alternative methods, i.e. K-Nearest Neighbors and One-class SVM, to estimate a
saliency score for each image patch, through which distinctive features stand
out without using identity labels in the training procedure. (3) saliency
matching is proposed based on patch matching. Matching patches with
inconsistent saliency brings penalty, and images of the same identity are
recognized by minimizing the saliency matching cost. (4) Furthermore, saliency
matching is tightly integrated with patch matching in a unified structural
RankSVM learning framework. The effectiveness of our approach is validated on
the VIPeR dataset and the CUHK01 dataset. Our approach outperforms the
state-of-the-art person re-identification methods on both datasets.
| no_new_dataset | 0.950778 |
1412.1947 | Aditya AV Sastry Mr. | Aditya AV Sastry and Kalyan Netti | A parallel sampling based clustering | null | null | null | null | cs.LG | http://creativecommons.org/licenses/publicdomain/ | The problem of automatically clustering data is an age old problem. People
have created numerous algorithms to tackle this problem. The execution time of
any of this algorithm grows with the number of input points and the number of
cluster centers required. To reduce the number of input points we could average
the points locally and use the means or the local centers as the input for
clustering. However since the required number of local centers is very high,
running the clustering algorithm on the entire dataset to obtain these
representational points is very time consuming. To remedy this problem, in this
paper we are proposing two subclustering schemes where by we subdivide the
dataset into smaller sets and run the clustering algorithm on the smaller
datasets to obtain the required number of datapoints to run our clustering
algorithm with. As we are subdividing the given dataset, we could run
clustering algorithm on each smaller piece of the dataset in parallel. We found
that both parallel and serial execution of this method to be much faster than
the original clustering algorithm and error in running the clustering algorithm
on a reduced set to be very less.
| [
{
"version": "v1",
"created": "Fri, 5 Dec 2014 10:50:31 GMT"
}
] | 2014-12-08T00:00:00 | [
[
"Sastry",
"Aditya AV",
""
],
[
"Netti",
"Kalyan",
""
]
] | TITLE: A parallel sampling based clustering
ABSTRACT: The problem of automatically clustering data is an age old problem. People
have created numerous algorithms to tackle this problem. The execution time of
any of this algorithm grows with the number of input points and the number of
cluster centers required. To reduce the number of input points we could average
the points locally and use the means or the local centers as the input for
clustering. However since the required number of local centers is very high,
running the clustering algorithm on the entire dataset to obtain these
representational points is very time consuming. To remedy this problem, in this
paper we are proposing two subclustering schemes where by we subdivide the
dataset into smaller sets and run the clustering algorithm on the smaller
datasets to obtain the required number of datapoints to run our clustering
algorithm with. As we are subdividing the given dataset, we could run
clustering algorithm on each smaller piece of the dataset in parallel. We found
that both parallel and serial execution of this method to be much faster than
the original clustering algorithm and error in running the clustering algorithm
on a reduced set to be very less.
| no_new_dataset | 0.95388 |
1409.1801 | Stephen Plaza | Stephen M. Plaza, Toufiq Parag, Gary B. Huang, Donald J. Olbris,
Mathew A. Saunders, Patricia K. Rivlin | Annotating Synapses in Large EM Datasets | null | null | null | null | q-bio.QM cs.CV q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reconstructing neuronal circuits at the level of synapses is a central
problem in neuroscience and becoming a focus of the emerging field of
connectomics. To date, electron microscopy (EM) is the most proven technique
for identifying and quantifying synaptic connections. As advances in EM make
acquiring larger datasets possible, subsequent manual synapse identification
({\em i.e.}, proofreading) for deciphering a connectome becomes a major time
bottleneck. Here we introduce a large-scale, high-throughput, and
semi-automated methodology to efficiently identify synapses. We successfully
applied our methodology to the Drosophila medulla optic lobe, annotating many
more synapses than previous connectome efforts. Our approaches are extensible
and will make the often complicated process of synapse identification
accessible to a wider-community of potential proofreaders.
| [
{
"version": "v1",
"created": "Fri, 5 Sep 2014 13:52:47 GMT"
},
{
"version": "v2",
"created": "Thu, 4 Dec 2014 16:18:01 GMT"
}
] | 2014-12-05T00:00:00 | [
[
"Plaza",
"Stephen M.",
""
],
[
"Parag",
"Toufiq",
""
],
[
"Huang",
"Gary B.",
""
],
[
"Olbris",
"Donald J.",
""
],
[
"Saunders",
"Mathew A.",
""
],
[
"Rivlin",
"Patricia K.",
""
]
] | TITLE: Annotating Synapses in Large EM Datasets
ABSTRACT: Reconstructing neuronal circuits at the level of synapses is a central
problem in neuroscience and becoming a focus of the emerging field of
connectomics. To date, electron microscopy (EM) is the most proven technique
for identifying and quantifying synaptic connections. As advances in EM make
acquiring larger datasets possible, subsequent manual synapse identification
({\em i.e.}, proofreading) for deciphering a connectome becomes a major time
bottleneck. Here we introduce a large-scale, high-throughput, and
semi-automated methodology to efficiently identify synapses. We successfully
applied our methodology to the Drosophila medulla optic lobe, annotating many
more synapses than previous connectome efforts. Our approaches are extensible
and will make the often complicated process of synapse identification
accessible to a wider-community of potential proofreaders.
| no_new_dataset | 0.951953 |
1412.1602 | Jan Chorowski | Jan Chorowski, Dzmitry Bahdanau, Kyunghyun Cho, Yoshua Bengio | End-to-end Continuous Speech Recognition using Attention-based Recurrent
NN: First Results | As accepted to: Deep Learning and Representation Learning Workshop,
NIPS 2014 | null | null | null | cs.NE cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We replace the Hidden Markov Model (HMM) which is traditionally used in in
continuous speech recognition with a bi-directional recurrent neural network
encoder coupled to a recurrent neural network decoder that directly emits a
stream of phonemes. The alignment between the input and output sequences is
established using an attention mechanism: the decoder emits each symbol based
on a context created with a subset of input symbols elected by the attention
mechanism. We report initial results demonstrating that this new approach
achieves phoneme error rates that are comparable to the state-of-the-art
HMM-based decoders, on the TIMIT dataset.
| [
{
"version": "v1",
"created": "Thu, 4 Dec 2014 10:00:19 GMT"
}
] | 2014-12-05T00:00:00 | [
[
"Chorowski",
"Jan",
""
],
[
"Bahdanau",
"Dzmitry",
""
],
[
"Cho",
"Kyunghyun",
""
],
[
"Bengio",
"Yoshua",
""
]
] | TITLE: End-to-end Continuous Speech Recognition using Attention-based Recurrent
NN: First Results
ABSTRACT: We replace the Hidden Markov Model (HMM) which is traditionally used in in
continuous speech recognition with a bi-directional recurrent neural network
encoder coupled to a recurrent neural network decoder that directly emits a
stream of phonemes. The alignment between the input and output sequences is
established using an attention mechanism: the decoder emits each symbol based
on a context created with a subset of input symbols elected by the attention
mechanism. We report initial results demonstrating that this new approach
achieves phoneme error rates that are comparable to the state-of-the-art
HMM-based decoders, on the TIMIT dataset.
| no_new_dataset | 0.953188 |
1412.1632 | Lei Yu | Lei Yu, Karl Moritz Hermann, Phil Blunsom and Stephen Pulman | Deep Learning for Answer Sentence Selection | 9 pages, accepted by NIPS deep learning workshop | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Answer sentence selection is the task of identifying sentences that contain
the answer to a given question. This is an important problem in its own right
as well as in the larger context of open domain question answering. We propose
a novel approach to solving this task via means of distributed representations,
and learn to match questions with answers by considering their semantic
encoding. This contrasts prior work on this task, which typically relies on
classifiers with large numbers of hand-crafted syntactic and semantic features
and various external resources. Our approach does not require any feature
engineering nor does it involve specialist linguistic data, making this model
easily applicable to a wide range of domains and languages. Experimental
results on a standard benchmark dataset from TREC demonstrate that---despite
its simplicity---our model matches state of the art performance on the answer
sentence selection task.
| [
{
"version": "v1",
"created": "Thu, 4 Dec 2014 11:53:02 GMT"
}
] | 2014-12-05T00:00:00 | [
[
"Yu",
"Lei",
""
],
[
"Hermann",
"Karl Moritz",
""
],
[
"Blunsom",
"Phil",
""
],
[
"Pulman",
"Stephen",
""
]
] | TITLE: Deep Learning for Answer Sentence Selection
ABSTRACT: Answer sentence selection is the task of identifying sentences that contain
the answer to a given question. This is an important problem in its own right
as well as in the larger context of open domain question answering. We propose
a novel approach to solving this task via means of distributed representations,
and learn to match questions with answers by considering their semantic
encoding. This contrasts prior work on this task, which typically relies on
classifiers with large numbers of hand-crafted syntactic and semantic features
and various external resources. Our approach does not require any feature
engineering nor does it involve specialist linguistic data, making this model
easily applicable to a wide range of domains and languages. Experimental
results on a standard benchmark dataset from TREC demonstrate that---despite
its simplicity---our model matches state of the art performance on the answer
sentence selection task.
| no_new_dataset | 0.946547 |
1412.1710 | Kaiming He | Kaiming He, Jian Sun | Convolutional Neural Networks at Constrained Time Cost | 8-page technical report | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Though recent advanced convolutional neural networks (CNNs) have been
improving the image recognition accuracy, the models are getting more complex
and time-consuming. For real-world applications in industrial and commercial
scenarios, engineers and developers are often faced with the requirement of
constrained time budget. In this paper, we investigate the accuracy of CNNs
under constrained time cost. Under this constraint, the designs of the network
architectures should exhibit as trade-offs among the factors like depth,
numbers of filters, filter sizes, etc. With a series of controlled comparisons,
we progressively modify a baseline model while preserving its time complexity.
This is also helpful for understanding the importance of the factors in network
designs. We present an architecture that achieves very competitive accuracy in
the ImageNet dataset (11.8% top-5 error, 10-view test), yet is 20% faster than
"AlexNet" (16.0% top-5 error, 10-view test).
| [
{
"version": "v1",
"created": "Thu, 4 Dec 2014 16:00:47 GMT"
}
] | 2014-12-05T00:00:00 | [
[
"He",
"Kaiming",
""
],
[
"Sun",
"Jian",
""
]
] | TITLE: Convolutional Neural Networks at Constrained Time Cost
ABSTRACT: Though recent advanced convolutional neural networks (CNNs) have been
improving the image recognition accuracy, the models are getting more complex
and time-consuming. For real-world applications in industrial and commercial
scenarios, engineers and developers are often faced with the requirement of
constrained time budget. In this paper, we investigate the accuracy of CNNs
under constrained time cost. Under this constraint, the designs of the network
architectures should exhibit as trade-offs among the factors like depth,
numbers of filters, filter sizes, etc. With a series of controlled comparisons,
we progressively modify a baseline model while preserving its time complexity.
This is also helpful for understanding the importance of the factors in network
designs. We present an architecture that achieves very competitive accuracy in
the ImageNet dataset (11.8% top-5 error, 10-view test), yet is 20% faster than
"AlexNet" (16.0% top-5 error, 10-view test).
| no_new_dataset | 0.951863 |
1412.1788 | Francis Bach | Felipe Yanez (LIENS, INRIA Paris - Rocquencourt), Francis Bach (LIENS,
INRIA Paris - Rocquencourt) | Primal-Dual Algorithms for Non-negative Matrix Factorization with the
Kullback-Leibler Divergence | null | null | null | null | cs.LG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Non-negative matrix factorization (NMF) approximates a given matrix as a
product of two non-negative matrices. Multiplicative algorithms deliver
reliable results, but they show slow convergence for high-dimensional data and
may be stuck away from local minima. Gradient descent methods have better
behavior, but only apply to smooth losses such as the least-squares loss. In
this article, we propose a first-order primal-dual algorithm for non-negative
decomposition problems (where one factor is fixed) with the KL divergence,
based on the Chambolle-Pock algorithm. All required computations may be
obtained in closed form and we provide an efficient heuristic way to select
step-sizes. By using alternating optimization, our algorithm readily extends to
NMF and, on synthetic examples, face recognition or music source separation
datasets, it is either faster than existing algorithms, or leads to improved
local optima, or both.
| [
{
"version": "v1",
"created": "Thu, 4 Dec 2014 19:56:02 GMT"
}
] | 2014-12-05T00:00:00 | [
[
"Yanez",
"Felipe",
"",
"LIENS, INRIA Paris - Rocquencourt"
],
[
"Bach",
"Francis",
"",
"LIENS,\n INRIA Paris - Rocquencourt"
]
] | TITLE: Primal-Dual Algorithms for Non-negative Matrix Factorization with the
Kullback-Leibler Divergence
ABSTRACT: Non-negative matrix factorization (NMF) approximates a given matrix as a
product of two non-negative matrices. Multiplicative algorithms deliver
reliable results, but they show slow convergence for high-dimensional data and
may be stuck away from local minima. Gradient descent methods have better
behavior, but only apply to smooth losses such as the least-squares loss. In
this article, we propose a first-order primal-dual algorithm for non-negative
decomposition problems (where one factor is fixed) with the KL divergence,
based on the Chambolle-Pock algorithm. All required computations may be
obtained in closed form and we provide an efficient heuristic way to select
step-sizes. By using alternating optimization, our algorithm readily extends to
NMF and, on synthetic examples, face recognition or music source separation
datasets, it is either faster than existing algorithms, or leads to improved
local optima, or both.
| no_new_dataset | 0.942771 |
1309.5643 | Veronika Cheplygina | Veronika Cheplygina, David M. J. Tax, and Marco Loog | Multiple Instance Learning with Bag Dissimilarities | Pattern Recognition, in press | Pattern Recognition 48.1 (2015): 264-275 | 10.1016/j.patcog.2014.07.022 | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multiple instance learning (MIL) is concerned with learning from sets (bags)
of objects (instances), where the individual instance labels are ambiguous. In
this setting, supervised learning cannot be applied directly. Often,
specialized MIL methods learn by making additional assumptions about the
relationship of the bag labels and instance labels. Such assumptions may fit a
particular dataset, but do not generalize to the whole range of MIL problems.
Other MIL methods shift the focus of assumptions from the labels to the overall
(dis)similarity of bags, and therefore learn from bags directly. We propose to
represent each bag by a vector of its dissimilarities to other bags in the
training set, and treat these dissimilarities as a feature representation. We
show several alternatives to define a dissimilarity between bags and discuss
which definitions are more suitable for particular MIL problems. The
experimental results show that the proposed approach is computationally
inexpensive, yet very competitive with state-of-the-art algorithms on a wide
range of MIL datasets.
| [
{
"version": "v1",
"created": "Sun, 22 Sep 2013 20:24:50 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Feb 2014 13:13:11 GMT"
},
{
"version": "v3",
"created": "Tue, 12 Aug 2014 09:04:32 GMT"
}
] | 2014-12-04T00:00:00 | [
[
"Cheplygina",
"Veronika",
""
],
[
"Tax",
"David M. J.",
""
],
[
"Loog",
"Marco",
""
]
] | TITLE: Multiple Instance Learning with Bag Dissimilarities
ABSTRACT: Multiple instance learning (MIL) is concerned with learning from sets (bags)
of objects (instances), where the individual instance labels are ambiguous. In
this setting, supervised learning cannot be applied directly. Often,
specialized MIL methods learn by making additional assumptions about the
relationship of the bag labels and instance labels. Such assumptions may fit a
particular dataset, but do not generalize to the whole range of MIL problems.
Other MIL methods shift the focus of assumptions from the labels to the overall
(dis)similarity of bags, and therefore learn from bags directly. We propose to
represent each bag by a vector of its dissimilarities to other bags in the
training set, and treat these dissimilarities as a feature representation. We
show several alternatives to define a dissimilarity between bags and discuss
which definitions are more suitable for particular MIL problems. The
experimental results show that the proposed approach is computationally
inexpensive, yet very competitive with state-of-the-art algorithms on a wide
range of MIL datasets.
| no_new_dataset | 0.947624 |
1412.1138 | Ben Fulcher | B. D. Fulcher, A. E. Georgieva, C. W. G. Redman, Nick S. Jones | Highly comparative fetal heart rate analysis | 7 pages, 4 figures | Fulcher, B. D., Georgieva, A., Redman, C. W., & Jones, N. S.
(2012). Highly comparative fetal heart rate analysis (pp. 3135-3138).
Presented at the 34th Annual International Conference of the IEEE EMBS, San
Diego, CA, USA | 10.1109/EMBC.2012.6346629 | null | cs.LG cs.AI q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A database of fetal heart rate (FHR) time series measured from 7221 patients
during labor is analyzed with the aim of learning the types of features of
these recordings that are informative of low cord pH. Our 'highly comparative'
analysis involves extracting over 9000 time-series analysis features from each
FHR time series, including measures of autocorrelation, entropy, distribution,
and various model fits. This diverse collection of features was developed in
previous work, and is publicly available. We describe five features that most
accurately classify a balanced training set of 59 'low pH' and 59 'normal pH'
FHR recordings. We then describe five of the features with the strongest linear
correlation to cord pH across the full dataset of FHR time series. The features
identified in this work may be used as part of a system for guiding
intervention during labor in future. This work successfully demonstrates the
utility of comparing across a large, interdisciplinary literature on
time-series analysis to automatically contribute new scientific results for
specific biomedical signal processing challenges.
| [
{
"version": "v1",
"created": "Wed, 3 Dec 2014 00:00:42 GMT"
}
] | 2014-12-04T00:00:00 | [
[
"Fulcher",
"B. D.",
""
],
[
"Georgieva",
"A. E.",
""
],
[
"Redman",
"C. W. G.",
""
],
[
"Jones",
"Nick S.",
""
]
] | TITLE: Highly comparative fetal heart rate analysis
ABSTRACT: A database of fetal heart rate (FHR) time series measured from 7221 patients
during labor is analyzed with the aim of learning the types of features of
these recordings that are informative of low cord pH. Our 'highly comparative'
analysis involves extracting over 9000 time-series analysis features from each
FHR time series, including measures of autocorrelation, entropy, distribution,
and various model fits. This diverse collection of features was developed in
previous work, and is publicly available. We describe five features that most
accurately classify a balanced training set of 59 'low pH' and 59 'normal pH'
FHR recordings. We then describe five of the features with the strongest linear
correlation to cord pH across the full dataset of FHR time series. The features
identified in this work may be used as part of a system for guiding
intervention during labor in future. This work successfully demonstrates the
utility of comparing across a large, interdisciplinary literature on
time-series analysis to automatically contribute new scientific results for
specific biomedical signal processing challenges.
| no_new_dataset | 0.929504 |
1412.1194 | Feng Shi | Feng Shi, Robert Laganiere, Emil Petriu | Gradient Boundary Histograms for Action Recognition | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a high efficient local spatiotemporal descriptor,
called gradient boundary histograms (GBH). The proposed GBH descriptor is built
on simple spatio-temporal gradients, which are fast to compute. We demonstrate
that it can better represent local structure and motion than other
gradient-based descriptors, and significantly outperforms them on large
realistic datasets. A comprehensive evaluation shows that the recognition
accuracy is preserved while the spatial resolution is greatly reduced, which
yields both high efficiency and low memory usage.
| [
{
"version": "v1",
"created": "Wed, 3 Dec 2014 05:23:03 GMT"
}
] | 2014-12-04T00:00:00 | [
[
"Shi",
"Feng",
""
],
[
"Laganiere",
"Robert",
""
],
[
"Petriu",
"Emil",
""
]
] | TITLE: Gradient Boundary Histograms for Action Recognition
ABSTRACT: This paper introduces a high efficient local spatiotemporal descriptor,
called gradient boundary histograms (GBH). The proposed GBH descriptor is built
on simple spatio-temporal gradients, which are fast to compute. We demonstrate
that it can better represent local structure and motion than other
gradient-based descriptors, and significantly outperforms them on large
realistic datasets. A comprehensive evaluation shows that the recognition
accuracy is preserved while the spatial resolution is greatly reduced, which
yields both high efficiency and low memory usage.
| no_new_dataset | 0.949669 |
1412.1442 | Maxwell Collins | Maxwell D. Collins and Pushmeet Kohli | Memory Bounded Deep Convolutional Networks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we investigate the use of sparsity-inducing regularizers during
training of Convolution Neural Networks (CNNs). These regularizers encourage
that fewer connections in the convolution and fully connected layers take
non-zero values and in effect result in sparse connectivity between hidden
units in the deep network. This in turn reduces the memory and runtime cost
involved in deploying the learned CNNs. We show that training with such
regularization can still be performed using stochastic gradient descent
implying that it can be used easily in existing codebases. Experimental
evaluation of our approach on MNIST, CIFAR, and ImageNet datasets shows that
our regularizers can result in dramatic reductions in memory requirements. For
instance, when applied on AlexNet, our method can reduce the memory consumption
by a factor of four with minimal loss in accuracy.
| [
{
"version": "v1",
"created": "Wed, 3 Dec 2014 19:08:38 GMT"
}
] | 2014-12-04T00:00:00 | [
[
"Collins",
"Maxwell D.",
""
],
[
"Kohli",
"Pushmeet",
""
]
] | TITLE: Memory Bounded Deep Convolutional Networks
ABSTRACT: In this work, we investigate the use of sparsity-inducing regularizers during
training of Convolution Neural Networks (CNNs). These regularizers encourage
that fewer connections in the convolution and fully connected layers take
non-zero values and in effect result in sparse connectivity between hidden
units in the deep network. This in turn reduces the memory and runtime cost
involved in deploying the learned CNNs. We show that training with such
regularization can still be performed using stochastic gradient descent
implying that it can be used easily in existing codebases. Experimental
evaluation of our approach on MNIST, CIFAR, and ImageNet datasets shows that
our regularizers can result in dramatic reductions in memory requirements. For
instance, when applied on AlexNet, our method can reduce the memory consumption
by a factor of four with minimal loss in accuracy.
| no_new_dataset | 0.949248 |
1409.3809 | Daniel Crankshaw | Daniel Crankshaw, Peter Bailis, Joseph E. Gonzalez, Haoyuan Li, Zhao
Zhang, Michael J. Franklin, Ali Ghodsi, Michael I. Jordan | The Missing Piece in Complex Analytics: Low Latency, Scalable Model
Management and Serving with Velox | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To support complex data-intensive applications such as personalized
recommendations, targeted advertising, and intelligent services, the data
management community has focused heavily on the design of systems to support
training complex models on large datasets. Unfortunately, the design of these
systems largely ignores a critical component of the overall analytics process:
the deployment and serving of models at scale. In this work, we present Velox,
a new component of the Berkeley Data Analytics Stack. Velox is a data
management system for facilitating the next steps in real-world, large-scale
analytics pipelines: online model management, maintenance, and serving. Velox
provides end-user applications and services with a low-latency, intuitive
interface to models, transforming the raw statistical models currently trained
using existing offline large-scale compute frameworks into full-blown,
end-to-end data products capable of recommending products, targeting
advertisements, and personalizing web content. To provide up-to-date results
for these complex models, Velox also facilitates lightweight online model
maintenance and selection (i.e., dynamic weighting). In this paper, we describe
the challenges and architectural considerations required to achieve this
functionality, including the abilities to span online and offline systems, to
adaptively adjust model materialization strategies, and to exploit inherent
statistical properties such as model error tolerance, all while operating at
"Big Data" scale.
| [
{
"version": "v1",
"created": "Fri, 12 Sep 2014 18:12:24 GMT"
},
{
"version": "v2",
"created": "Mon, 1 Dec 2014 23:20:30 GMT"
}
] | 2014-12-03T00:00:00 | [
[
"Crankshaw",
"Daniel",
""
],
[
"Bailis",
"Peter",
""
],
[
"Gonzalez",
"Joseph E.",
""
],
[
"Li",
"Haoyuan",
""
],
[
"Zhang",
"Zhao",
""
],
[
"Franklin",
"Michael J.",
""
],
[
"Ghodsi",
"Ali",
""
],
[
"Jordan",
"Michael I.",
""
]
] | TITLE: The Missing Piece in Complex Analytics: Low Latency, Scalable Model
Management and Serving with Velox
ABSTRACT: To support complex data-intensive applications such as personalized
recommendations, targeted advertising, and intelligent services, the data
management community has focused heavily on the design of systems to support
training complex models on large datasets. Unfortunately, the design of these
systems largely ignores a critical component of the overall analytics process:
the deployment and serving of models at scale. In this work, we present Velox,
a new component of the Berkeley Data Analytics Stack. Velox is a data
management system for facilitating the next steps in real-world, large-scale
analytics pipelines: online model management, maintenance, and serving. Velox
provides end-user applications and services with a low-latency, intuitive
interface to models, transforming the raw statistical models currently trained
using existing offline large-scale compute frameworks into full-blown,
end-to-end data products capable of recommending products, targeting
advertisements, and personalizing web content. To provide up-to-date results
for these complex models, Velox also facilitates lightweight online model
maintenance and selection (i.e., dynamic weighting). In this paper, we describe
the challenges and architectural considerations required to achieve this
functionality, including the abilities to span online and offline systems, to
adaptively adjust model materialization strategies, and to exploit inherent
statistical properties such as model error tolerance, all while operating at
"Big Data" scale.
| no_new_dataset | 0.946597 |
1402.3384 | Weina Wang | Weina Wang, Lei Ying and Junshan Zhang | A Minimax Distortion View of Differentially Private Query Release | null | null | null | null | cs.CR cs.DB cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of differentially private query release through a
synthetic database approach. Departing from the existing approaches that
require the query set to be specified in advance, we advocate to devise
query-set independent mechanisms, with an ambitious goal of providing accurate
answers, while meeting the privacy constraints, for all queries in a general
query class. Specifically, a differentially private mechanism is constructed to
"encode" rich stochastic structure into the synthetic database, and
"customized" companion estimators are then derived to provide accurate answers
by making use of all available information, including the mechanism (which is
public information) and the query functions. Accordingly, the distortion under
the best of this kind of mechanisms at the worst-case query in a general query
class, so called the minimax distortion, provides a fundamental
characterization of differentially private query release.
For the general class of statistical queries, we prove that with the
squared-error distortion measure, the minimax distortion is $O(1/n)$ by
deriving asymptotically tight upper and lower bounds in the regime that the
database size $n$ goes to infinity. The upper bound is achievable by a
mechanism $\mathcal{E}$ and its corresponding companion estimators, which
points directly to the feasibility of the proposed approach in large databases.
We further evaluate the mechanism $\mathcal{E}$ and the companion estimators
through experiments on real datasets from Netflix and Facebook. Experimental
results show improvement over the state-of-art MWEM algorithm and verify the
scaling behavior $O(1/n)$ of the minimax distortion.
| [
{
"version": "v1",
"created": "Fri, 14 Feb 2014 06:54:49 GMT"
},
{
"version": "v2",
"created": "Mon, 1 Dec 2014 07:01:08 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Wang",
"Weina",
""
],
[
"Ying",
"Lei",
""
],
[
"Zhang",
"Junshan",
""
]
] | TITLE: A Minimax Distortion View of Differentially Private Query Release
ABSTRACT: We consider the problem of differentially private query release through a
synthetic database approach. Departing from the existing approaches that
require the query set to be specified in advance, we advocate to devise
query-set independent mechanisms, with an ambitious goal of providing accurate
answers, while meeting the privacy constraints, for all queries in a general
query class. Specifically, a differentially private mechanism is constructed to
"encode" rich stochastic structure into the synthetic database, and
"customized" companion estimators are then derived to provide accurate answers
by making use of all available information, including the mechanism (which is
public information) and the query functions. Accordingly, the distortion under
the best of this kind of mechanisms at the worst-case query in a general query
class, so called the minimax distortion, provides a fundamental
characterization of differentially private query release.
For the general class of statistical queries, we prove that with the
squared-error distortion measure, the minimax distortion is $O(1/n)$ by
deriving asymptotically tight upper and lower bounds in the regime that the
database size $n$ goes to infinity. The upper bound is achievable by a
mechanism $\mathcal{E}$ and its corresponding companion estimators, which
points directly to the feasibility of the proposed approach in large databases.
We further evaluate the mechanism $\mathcal{E}$ and the companion estimators
through experiments on real datasets from Netflix and Facebook. Experimental
results show improvement over the state-of-art MWEM algorithm and verify the
scaling behavior $O(1/n)$ of the minimax distortion.
| no_new_dataset | 0.944228 |
1410.3726 | Chee Seng Chan | Chern Hong Lim, Anhar Risnumawan and Chee Seng Chan | Scene Image is Non-Mutually Exclusive - A Fuzzy Qualitative Scene
Understanding | Accepted in IEEE Transactions on Fuzzy Systems | IEEE Transactions on Fuzzy Systems, vol. 22(6), pp. 1541 - 1556,
2014 | 10.1109/TFUZZ.2014.2298233 | null | cs.CV cs.AI cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ambiguity or uncertainty is a pervasive element of many real world decision
making processes. Variation in decisions is a norm in this situation when the
same problem is posed to different subjects. Psychological and metaphysical
research had proven that decision making by human is subjective. It is
influenced by many factors such as experience, age, background, etc. Scene
understanding is one of the computer vision problems that fall into this
category. Conventional methods relax this problem by assuming scene images are
mutually exclusive; and therefore, focus on developing different approaches to
perform the binary classification tasks. In this paper, we show that scene
images are non-mutually exclusive, and propose the Fuzzy Qualitative Rank
Classifier (FQRC) to tackle the aforementioned problems. The proposed FQRC
provides a ranking interpretation instead of binary decision. Evaluations in
term of qualitative and quantitative using large numbers and challenging public
scene datasets have shown the effectiveness of our proposed method in modeling
the non-mutually exclusive scene images.
| [
{
"version": "v1",
"created": "Tue, 14 Oct 2014 15:19:43 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Lim",
"Chern Hong",
""
],
[
"Risnumawan",
"Anhar",
""
],
[
"Chan",
"Chee Seng",
""
]
] | TITLE: Scene Image is Non-Mutually Exclusive - A Fuzzy Qualitative Scene
Understanding
ABSTRACT: Ambiguity or uncertainty is a pervasive element of many real world decision
making processes. Variation in decisions is a norm in this situation when the
same problem is posed to different subjects. Psychological and metaphysical
research had proven that decision making by human is subjective. It is
influenced by many factors such as experience, age, background, etc. Scene
understanding is one of the computer vision problems that fall into this
category. Conventional methods relax this problem by assuming scene images are
mutually exclusive; and therefore, focus on developing different approaches to
perform the binary classification tasks. In this paper, we show that scene
images are non-mutually exclusive, and propose the Fuzzy Qualitative Rank
Classifier (FQRC) to tackle the aforementioned problems. The proposed FQRC
provides a ranking interpretation instead of binary decision. Evaluations in
term of qualitative and quantitative using large numbers and challenging public
scene datasets have shown the effectiveness of our proposed method in modeling
the non-mutually exclusive scene images.
| no_new_dataset | 0.957675 |
1412.0008 | Robert Templeman | Mohammed Korayem, Robert Templeman, Dennis Chen, David Crandall, Apu
Kapadia | ScreenAvoider: Protecting Computer Screens from Ubiquitous Cameras | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We live and work in environments that are inundated with cameras embedded in
devices such as phones, tablets, laptops, and monitors. Newer wearable devices
like Google Glass, Narrative Clip, and Autographer offer the ability to quietly
log our lives with cameras from a `first person' perspective. While capturing
several meaningful and interesting moments, a significant number of images
captured by these wearable cameras can contain computer screens. Given the
potentially sensitive information that is visible on our displays, there is a
need to guard computer screens from undesired photography. People need
protection against photography of their screens, whether by other people's
cameras or their own cameras.
We present ScreenAvoider, a framework that controls the collection and
disclosure of images with computer screens and their sensitive content.
ScreenAvoider can detect images with computer screens with high accuracy and
can even go so far as to discriminate amongst screen content. We also introduce
a ScreenTag system that aids in the identification of screen content, flagging
images with highly sensitive content such as messaging applications or email
webpages. We evaluate our concept on realistic lifelogging datasets, showing
that ScreenAvoider provides a practical and useful solution that can help users
manage their privacy.
| [
{
"version": "v1",
"created": "Fri, 28 Nov 2014 01:50:53 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Korayem",
"Mohammed",
""
],
[
"Templeman",
"Robert",
""
],
[
"Chen",
"Dennis",
""
],
[
"Crandall",
"David",
""
],
[
"Kapadia",
"Apu",
""
]
] | TITLE: ScreenAvoider: Protecting Computer Screens from Ubiquitous Cameras
ABSTRACT: We live and work in environments that are inundated with cameras embedded in
devices such as phones, tablets, laptops, and monitors. Newer wearable devices
like Google Glass, Narrative Clip, and Autographer offer the ability to quietly
log our lives with cameras from a `first person' perspective. While capturing
several meaningful and interesting moments, a significant number of images
captured by these wearable cameras can contain computer screens. Given the
potentially sensitive information that is visible on our displays, there is a
need to guard computer screens from undesired photography. People need
protection against photography of their screens, whether by other people's
cameras or their own cameras.
We present ScreenAvoider, a framework that controls the collection and
disclosure of images with computer screens and their sensitive content.
ScreenAvoider can detect images with computer screens with high accuracy and
can even go so far as to discriminate amongst screen content. We also introduce
a ScreenTag system that aids in the identification of screen content, flagging
images with highly sensitive content such as messaging applications or email
webpages. We evaluate our concept on realistic lifelogging datasets, showing
that ScreenAvoider provides a practical and useful solution that can help users
manage their privacy.
| no_new_dataset | 0.916559 |
1412.0059 | Takashi Sakuragawa | Kent Miyajima and Takashi Sakuragawa | Continuous and robust clustering coefficients for weighted and directed
networks | 29 pages, 14 figures | null | null | null | physics.soc-ph cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce new clustering coefficients for weighted networks. They are
continuous and robust against edge weight changes. Recently, generalized
clustering coefficients for weighted and directed networks have been proposed.
These generalizations have a common property, that their values are not
continuous. They are sensitive with edge weight changes, especially at zero
weight. With these generalizations, if vanishingly low weights of edges are
truncated to weight zero for some reason, the coefficient value may change
significantly from the original value. It is preferable that small changes of
edge weights cause small changes of coefficient value. We call this property
the continuity of generalized clustering coefficients. Our new coefficients
admit this property. In the past, few studies have focused on the continuity of
generalized clustering coefficients. In experiments, we performed comparative
assessments of existing and our generalizations. In the case of a real world
network dataset (C. Elegans Neural network), after adding random edge weight
errors, though the value of one discontinuous generalization was changed about
436%, the value of proposed one was only changed 0.2%.
| [
{
"version": "v1",
"created": "Sat, 29 Nov 2014 01:57:08 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Miyajima",
"Kent",
""
],
[
"Sakuragawa",
"Takashi",
""
]
] | TITLE: Continuous and robust clustering coefficients for weighted and directed
networks
ABSTRACT: We introduce new clustering coefficients for weighted networks. They are
continuous and robust against edge weight changes. Recently, generalized
clustering coefficients for weighted and directed networks have been proposed.
These generalizations have a common property, that their values are not
continuous. They are sensitive with edge weight changes, especially at zero
weight. With these generalizations, if vanishingly low weights of edges are
truncated to weight zero for some reason, the coefficient value may change
significantly from the original value. It is preferable that small changes of
edge weights cause small changes of coefficient value. We call this property
the continuity of generalized clustering coefficients. Our new coefficients
admit this property. In the past, few studies have focused on the continuity of
generalized clustering coefficients. In experiments, we performed comparative
assessments of existing and our generalizations. In the case of a real world
network dataset (C. Elegans Neural network), after adding random edge weight
errors, though the value of one discontinuous generalization was changed about
436%, the value of proposed one was only changed 0.2%.
| no_new_dataset | 0.949623 |
1412.0065 | Gr\'egory Rogez | Gregory Rogez, James S. Supancic III, Maryam Khademi, Jose Maria
Martinez Montiel, Deva Ramanan | 3D Hand Pose Detection in Egocentric RGB-D Images | 14 pages, 15 figures, extended version of the corresponding ECCV
workshop paper, submitted to International Journal of Computer Vision | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We focus on the task of everyday hand pose estimation from egocentric
viewpoints. For this task, we show that depth sensors are particularly
informative for extracting near-field interactions of the camera wearer with
his/her environment. Despite the recent advances in full-body pose estimation
using Kinect-like sensors, reliable monocular hand pose estimation in RGB-D
images is still an unsolved problem. The problem is considerably exacerbated
when analyzing hands performing daily activities from a first-person viewpoint,
due to severe occlusions arising from object manipulations and a limited
field-of-view. Our system addresses these difficulties by exploiting strong
priors over viewpoint and pose in a discriminative tracking-by-detection
framework. Our priors are operationalized through a photorealistic synthetic
model of egocentric scenes, which is used to generate training data for
learning depth-based pose classifiers. We evaluate our approach on an annotated
dataset of real egocentric object manipulation scenes and compare to both
commercial and academic approaches. Our method provides state-of-the-art
performance for both hand detection and pose estimation in egocentric RGB-D
images.
| [
{
"version": "v1",
"created": "Sat, 29 Nov 2014 03:19:56 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Rogez",
"Gregory",
""
],
[
"Supancic",
"James S.",
"III"
],
[
"Khademi",
"Maryam",
""
],
[
"Montiel",
"Jose Maria Martinez",
""
],
[
"Ramanan",
"Deva",
""
]
] | TITLE: 3D Hand Pose Detection in Egocentric RGB-D Images
ABSTRACT: We focus on the task of everyday hand pose estimation from egocentric
viewpoints. For this task, we show that depth sensors are particularly
informative for extracting near-field interactions of the camera wearer with
his/her environment. Despite the recent advances in full-body pose estimation
using Kinect-like sensors, reliable monocular hand pose estimation in RGB-D
images is still an unsolved problem. The problem is considerably exacerbated
when analyzing hands performing daily activities from a first-person viewpoint,
due to severe occlusions arising from object manipulations and a limited
field-of-view. Our system addresses these difficulties by exploiting strong
priors over viewpoint and pose in a discriminative tracking-by-detection
framework. Our priors are operationalized through a photorealistic synthetic
model of egocentric scenes, which is used to generate training data for
learning depth-based pose classifiers. We evaluate our approach on an annotated
dataset of real egocentric object manipulation scenes and compare to both
commercial and academic approaches. Our method provides state-of-the-art
performance for both hand detection and pose estimation in egocentric RGB-D
images.
| no_new_dataset | 0.946941 |
1412.0069 | Yonglong Tian | Yonglong Tian, Ping Luo, Xiaogang Wang, Xiaoou Tang | Pedestrian Detection aided by Deep Learning Semantic Tasks | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning methods have achieved great success in pedestrian detection,
owing to its ability to learn features from raw pixels. However, they mainly
capture middle-level representations, such as pose of pedestrian, but confuse
positive with hard negative samples, which have large ambiguity, e.g. the shape
and appearance of `tree trunk' or `wire pole' are similar to pedestrian in
certain viewpoint. This ambiguity can be distinguished by high-level
representation. To this end, this work jointly optimizes pedestrian detection
with semantic tasks, including pedestrian attributes (e.g. `carrying backpack')
and scene attributes (e.g. `road', `tree', and `horizontal'). Rather than
expensively annotating scene attributes, we transfer attributes information
from existing scene segmentation datasets to the pedestrian dataset, by
proposing a novel deep model to learn high-level features from multiple tasks
and multiple data sources. Since distinct tasks have distinct convergence rates
and data from different datasets have different distributions, a multi-task
objective function is carefully designed to coordinate tasks and reduce
discrepancies among datasets. The importance coefficients of tasks and network
parameters in this objective function can be iteratively estimated. Extensive
evaluations show that the proposed approach outperforms the state-of-the-art on
the challenging Caltech and ETH datasets, where it reduces the miss rates of
previous deep models by 17 and 5.5 percent, respectively.
| [
{
"version": "v1",
"created": "Sat, 29 Nov 2014 04:34:23 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Tian",
"Yonglong",
""
],
[
"Luo",
"Ping",
""
],
[
"Wang",
"Xiaogang",
""
],
[
"Tang",
"Xiaoou",
""
]
] | TITLE: Pedestrian Detection aided by Deep Learning Semantic Tasks
ABSTRACT: Deep learning methods have achieved great success in pedestrian detection,
owing to its ability to learn features from raw pixels. However, they mainly
capture middle-level representations, such as pose of pedestrian, but confuse
positive with hard negative samples, which have large ambiguity, e.g. the shape
and appearance of `tree trunk' or `wire pole' are similar to pedestrian in
certain viewpoint. This ambiguity can be distinguished by high-level
representation. To this end, this work jointly optimizes pedestrian detection
with semantic tasks, including pedestrian attributes (e.g. `carrying backpack')
and scene attributes (e.g. `road', `tree', and `horizontal'). Rather than
expensively annotating scene attributes, we transfer attributes information
from existing scene segmentation datasets to the pedestrian dataset, by
proposing a novel deep model to learn high-level features from multiple tasks
and multiple data sources. Since distinct tasks have distinct convergence rates
and data from different datasets have different distributions, a multi-task
objective function is carefully designed to coordinate tasks and reduce
discrepancies among datasets. The importance coefficients of tasks and network
parameters in this objective function can be iteratively estimated. Extensive
evaluations show that the proposed approach outperforms the state-of-the-art on
the challenging Caltech and ETH datasets, where it reduces the miss rates of
previous deep models by 17 and 5.5 percent, respectively.
| no_new_dataset | 0.941385 |
1412.0100 | Stefan Mathe | Stefan Mathe, Cristian Sminchisescu | Multiple Instance Reinforcement Learning for Efficient Weakly-Supervised
Detection in Images | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | State-of-the-art visual recognition and detection systems increasingly rely
on large amounts of training data and complex classifiers. Therefore it becomes
increasingly expensive both to manually annotate datasets and to keep running
times at levels acceptable for practical applications. In this paper, we
propose two solutions to address these issues. First, we introduce a weakly
supervised, segmentation-based approach to learn accurate detectors and image
classifiers from weak supervisory signals that provide only approximate
constraints on target localization. We illustrate our system on the problem of
action detection in static images (Pascal VOC Actions 2012), using human visual
search patterns as our training signal. Second, inspired from the
saccade-and-fixate operating principle of the human visual system, we use
reinforcement learning techniques to train efficient search models for
detection. Our sequential method is weakly supervised and general (it does not
require eye movements), finds optimal search strategies for any given detection
confidence function and achieves performance similar to exhaustive sliding
window search at a fraction of its computational cost.
| [
{
"version": "v1",
"created": "Sat, 29 Nov 2014 12:18:14 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Mathe",
"Stefan",
""
],
[
"Sminchisescu",
"Cristian",
""
]
] | TITLE: Multiple Instance Reinforcement Learning for Efficient Weakly-Supervised
Detection in Images
ABSTRACT: State-of-the-art visual recognition and detection systems increasingly rely
on large amounts of training data and complex classifiers. Therefore it becomes
increasingly expensive both to manually annotate datasets and to keep running
times at levels acceptable for practical applications. In this paper, we
propose two solutions to address these issues. First, we introduce a weakly
supervised, segmentation-based approach to learn accurate detectors and image
classifiers from weak supervisory signals that provide only approximate
constraints on target localization. We illustrate our system on the problem of
action detection in static images (Pascal VOC Actions 2012), using human visual
search patterns as our training signal. Second, inspired from the
saccade-and-fixate operating principle of the human visual system, we use
reinforcement learning techniques to train efficient search models for
detection. Our sequential method is weakly supervised and general (it does not
require eye movements), finds optimal search strategies for any given detection
confidence function and achieves performance similar to exhaustive sliding
window search at a fraction of its computational cost.
| no_new_dataset | 0.946646 |
1412.0296 | George Papandreou | George Papandreou and Iasonas Kokkinos and Pierre-Andr\'e Savalle | Untangling Local and Global Deformations in Deep Convolutional Networks
for Image Classification and Sliding Window Detection | 13 pages, 7 figures, 5 tables. arXiv admin note: substantial text
overlap with arXiv:1406.2732 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep Convolutional Neural Networks (DCNNs) commonly use generic `max-pooling'
(MP) layers to extract deformation-invariant features, but we argue in favor of
a more refined treatment. First, we introduce epitomic convolution as a
building block alternative to the common convolution-MP cascade of DCNNs; while
having identical complexity to MP, Epitomic Convolution allows for parameter
sharing across different filters, resulting in faster convergence and better
generalization. Second, we introduce a Multiple Instance Learning approach to
explicitly accommodate global translation and scaling when training a DCNN
exclusively with class labels. For this we rely on a `patchwork' data structure
that efficiently lays out all image scales and positions as candidates to a
DCNN. Factoring global and local deformations allows a DCNN to `focus its
resources' on the treatment of non-rigid deformations and yields a substantial
classification accuracy improvement. Third, further pursuing this idea, we
develop an efficient DCNN sliding window object detector that employs explicit
search over position, scale, and aspect ratio. We provide competitive image
classification and localization results on the ImageNet dataset and object
detection results on the Pascal VOC 2007 benchmark.
| [
{
"version": "v1",
"created": "Sun, 30 Nov 2014 22:20:17 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Papandreou",
"George",
""
],
[
"Kokkinos",
"Iasonas",
""
],
[
"Savalle",
"Pierre-André",
""
]
] | TITLE: Untangling Local and Global Deformations in Deep Convolutional Networks
for Image Classification and Sliding Window Detection
ABSTRACT: Deep Convolutional Neural Networks (DCNNs) commonly use generic `max-pooling'
(MP) layers to extract deformation-invariant features, but we argue in favor of
a more refined treatment. First, we introduce epitomic convolution as a
building block alternative to the common convolution-MP cascade of DCNNs; while
having identical complexity to MP, Epitomic Convolution allows for parameter
sharing across different filters, resulting in faster convergence and better
generalization. Second, we introduce a Multiple Instance Learning approach to
explicitly accommodate global translation and scaling when training a DCNN
exclusively with class labels. For this we rely on a `patchwork' data structure
that efficiently lays out all image scales and positions as candidates to a
DCNN. Factoring global and local deformations allows a DCNN to `focus its
resources' on the treatment of non-rigid deformations and yields a substantial
classification accuracy improvement. Third, further pursuing this idea, we
develop an efficient DCNN sliding window object detector that employs explicit
search over position, scale, and aspect ratio. We provide competitive image
classification and localization results on the ImageNet dataset and object
detection results on the Pascal VOC 2007 benchmark.
| no_new_dataset | 0.949716 |
1412.0327 | Jiajun Liu | Jiajun Liu, Kun Zhao, Saeed Khan, Mark Cameron, Raja Jurdak | Multi-scale Population and Mobility Estimation with Geo-tagged Tweets | 1st International Workshop on Big Data Analytics for Biosecurity
(BioBAD2015), 4 pages | null | null | null | cs.SI cs.CY physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent outbreaks of Ebola and Dengue viruses have again elevated the
significance of the capability to quickly predict disease spread in an emergent
situation. However, existing approaches usually rely heavily on the
time-consuming census processes, or the privacy-sensitive call logs, leading to
their unresponsive nature when facing the abruptly changing dynamics in the
event of an outbreak. In this paper we study the feasibility of using
large-scale Twitter data as a proxy of human mobility to model and predict
disease spread. We report that for Australia, Twitter users' distribution
correlates well the census-based population distribution, and that the Twitter
users' travel patterns appear to loosely follow the gravity law at multiple
scales of geographic distances, i.e. national level, state level and
metropolitan level. The radiation model is also evaluated on this dataset
though it has shown inferior fitness as a result of Australia's sparse
population and large landmass. The outcomes of the study form the cornerstones
for future work towards a model-based, responsive prediction method from
Twitter data for disease spread.
| [
{
"version": "v1",
"created": "Mon, 1 Dec 2014 01:48:53 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Liu",
"Jiajun",
""
],
[
"Zhao",
"Kun",
""
],
[
"Khan",
"Saeed",
""
],
[
"Cameron",
"Mark",
""
],
[
"Jurdak",
"Raja",
""
]
] | TITLE: Multi-scale Population and Mobility Estimation with Geo-tagged Tweets
ABSTRACT: Recent outbreaks of Ebola and Dengue viruses have again elevated the
significance of the capability to quickly predict disease spread in an emergent
situation. However, existing approaches usually rely heavily on the
time-consuming census processes, or the privacy-sensitive call logs, leading to
their unresponsive nature when facing the abruptly changing dynamics in the
event of an outbreak. In this paper we study the feasibility of using
large-scale Twitter data as a proxy of human mobility to model and predict
disease spread. We report that for Australia, Twitter users' distribution
correlates well the census-based population distribution, and that the Twitter
users' travel patterns appear to loosely follow the gravity law at multiple
scales of geographic distances, i.e. national level, state level and
metropolitan level. The radiation model is also evaluated on this dataset
though it has shown inferior fitness as a result of Australia's sparse
population and large landmass. The outcomes of the study form the cornerstones
for future work towards a model-based, responsive prediction method from
Twitter data for disease spread.
| no_new_dataset | 0.943608 |
1412.0488 | Pavel Tomancak | Tobias Pietzsch and Stephan Saalfeld and Stephan Preibisch and Pavel
Tomancak | BigDataViewer: Interactive Visualization and Image Processing for
Terabyte Data Sets | 38 pages, 1 main figure, 27 supplementary figures, under review at
Nature Methods | null | null | null | q-bio.QM cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasingly popular light sheet microscopy techniques generate very
large 3D time-lapse recordings of living biological specimen. The necessity to
make large volumetric datasets available for interactive visualization and
analysis has been widely recognized. However, existing solutions build on
dedicated servers to generate virtual slices that are transferred to the client
applications, practically leading to insufficient frame rates (less than 10
frames per second) for truly interactive experience. An easily accessible open
source solution for interactive arbitrary virtual re-slicing of very large
volumes and time series of volumes has yet been missing. We fill this gap with
BigDataViewer, a Fiji plugin to interactively navigate and visualize large
image sequences from both local and remote data sources.
| [
{
"version": "v1",
"created": "Mon, 1 Dec 2014 14:24:44 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Pietzsch",
"Tobias",
""
],
[
"Saalfeld",
"Stephan",
""
],
[
"Preibisch",
"Stephan",
""
],
[
"Tomancak",
"Pavel",
""
]
] | TITLE: BigDataViewer: Interactive Visualization and Image Processing for
Terabyte Data Sets
ABSTRACT: The increasingly popular light sheet microscopy techniques generate very
large 3D time-lapse recordings of living biological specimen. The necessity to
make large volumetric datasets available for interactive visualization and
analysis has been widely recognized. However, existing solutions build on
dedicated servers to generate virtual slices that are transferred to the client
applications, practically leading to insufficient frame rates (less than 10
frames per second) for truly interactive experience. An easily accessible open
source solution for interactive arbitrary virtual re-slicing of very large
volumes and time series of volumes has yet been missing. We fill this gap with
BigDataViewer, a Fiji plugin to interactively navigate and visualize large
image sequences from both local and remote data sources.
| no_new_dataset | 0.93744 |
1412.0630 | Sean Anderson | Sean Anderson, Timothy D. Barfoot, Chi Hay Tong, Simo S\"arkk\"a | Batch Nonlinear Continuous-Time Trajectory Estimation as Exactly Sparse
Gaussian Process Regression | Submitted to Autonomous Robots on 20 November 2014, manuscript #
AURO-D-14-00185, 16 pages, 7 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we revisit batch state estimation through the lens of Gaussian
process (GP) regression. We consider continuous-discrete estimation problems
wherein a trajectory is viewed as a one-dimensional GP, with time as the
independent variable. Our continuous-time prior can be defined by any
nonlinear, time-varying stochastic differential equation driven by white noise;
this allows the possibility of smoothing our trajectory estimates using a
variety of vehicle dynamics models (e.g., `constant-velocity'). We show that
this class of prior results in an inverse kernel matrix (i.e., covariance
matrix between all pairs of measurement times) that is exactly sparse
(block-tridiagonal) and that this can be exploited to carry out GP regression
(and interpolation) very efficiently. When the prior is based on a linear,
time-varying stochastic differential equation and the measurement model is also
linear, this GP approach is equivalent to classical, discrete-time smoothing
(at the measurement times); when a nonlinearity is present, we iterate over the
whole trajectory to maximize accuracy. We test the approach experimentally on a
simultaneous trajectory estimation and mapping problem using a mobile robot
dataset.
| [
{
"version": "v1",
"created": "Mon, 1 Dec 2014 20:24:08 GMT"
}
] | 2014-12-02T00:00:00 | [
[
"Anderson",
"Sean",
""
],
[
"Barfoot",
"Timothy D.",
""
],
[
"Tong",
"Chi Hay",
""
],
[
"Särkkä",
"Simo",
""
]
] | TITLE: Batch Nonlinear Continuous-Time Trajectory Estimation as Exactly Sparse
Gaussian Process Regression
ABSTRACT: In this paper, we revisit batch state estimation through the lens of Gaussian
process (GP) regression. We consider continuous-discrete estimation problems
wherein a trajectory is viewed as a one-dimensional GP, with time as the
independent variable. Our continuous-time prior can be defined by any
nonlinear, time-varying stochastic differential equation driven by white noise;
this allows the possibility of smoothing our trajectory estimates using a
variety of vehicle dynamics models (e.g., `constant-velocity'). We show that
this class of prior results in an inverse kernel matrix (i.e., covariance
matrix between all pairs of measurement times) that is exactly sparse
(block-tridiagonal) and that this can be exploited to carry out GP regression
(and interpolation) very efficiently. When the prior is based on a linear,
time-varying stochastic differential equation and the measurement model is also
linear, this GP approach is equivalent to classical, discrete-time smoothing
(at the measurement times); when a nonlinearity is present, we iterate over the
whole trajectory to maximize accuracy. We test the approach experimentally on a
simultaneous trajectory estimation and mapping problem using a mobile robot
dataset.
| no_new_dataset | 0.94887 |
1411.7441 | Stefano Ermon | Stefano Ermon, Ronan Le Bras, Santosh K. Suram, John M. Gregoire,
Carla Gomes, Bart Selman, Robert B. van Dover | Pattern Decomposition with Complex Combinatorial Constraints:
Application to Materials Discovery | null | null | null | null | cs.AI cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Identifying important components or factors in large amounts of noisy data is
a key problem in machine learning and data mining. Motivated by a pattern
decomposition problem in materials discovery, aimed at discovering new
materials for renewable energy, e.g. for fuel and solar cells, we introduce
CombiFD, a framework for factor based pattern decomposition that allows the
incorporation of a-priori knowledge as constraints, including complex
combinatorial constraints. In addition, we propose a new pattern decomposition
algorithm, called AMIQO, based on solving a sequence of (mixed-integer)
quadratic programs. Our approach considerably outperforms the state of the art
on the materials discovery problem, scaling to larger datasets and recovering
more precise and physically meaningful decompositions. We also show the
effectiveness of our approach for enforcing background knowledge on other
application domains.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 02:31:41 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Ermon",
"Stefano",
""
],
[
"Bras",
"Ronan Le",
""
],
[
"Suram",
"Santosh K.",
""
],
[
"Gregoire",
"John M.",
""
],
[
"Gomes",
"Carla",
""
],
[
"Selman",
"Bart",
""
],
[
"van Dover",
"Robert B.",
""
]
] | TITLE: Pattern Decomposition with Complex Combinatorial Constraints:
Application to Materials Discovery
ABSTRACT: Identifying important components or factors in large amounts of noisy data is
a key problem in machine learning and data mining. Motivated by a pattern
decomposition problem in materials discovery, aimed at discovering new
materials for renewable energy, e.g. for fuel and solar cells, we introduce
CombiFD, a framework for factor based pattern decomposition that allows the
incorporation of a-priori knowledge as constraints, including complex
combinatorial constraints. In addition, we propose a new pattern decomposition
algorithm, called AMIQO, based on solving a sequence of (mixed-integer)
quadratic programs. Our approach considerably outperforms the state of the art
on the materials discovery problem, scaling to larger datasets and recovering
more precise and physically meaningful decompositions. We also show the
effectiveness of our approach for enforcing background knowledge on other
application domains.
| no_new_dataset | 0.942665 |
1411.7445 | Chao Xu | Tao Han, Chao Xu, Ryan Loxton, Lei Xie | Bi-objective Optimization for Robust RGB-D Visual Odometry | null | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers a new bi-objective optimization formulation for robust
RGB-D visual odometry. We investigate two methods for solving the proposed
bi-objective optimization problem: the weighted sum method (in which the
objective functions are combined into a single objective function) and the
bounded objective method (in which one of the objective functions is optimized
and the value of the other objective function is bounded via a constraint). Our
experimental results for the open source TUM RGB-D dataset show that the new
bi-objective optimization formulation is superior to several existing RGB-D
odometry methods. In particular, the new formulation yields more accurate
motion estimates and is more robust when textural or structural features in the
image sequence are lacking.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 02:37:41 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Han",
"Tao",
""
],
[
"Xu",
"Chao",
""
],
[
"Loxton",
"Ryan",
""
],
[
"Xie",
"Lei",
""
]
] | TITLE: Bi-objective Optimization for Robust RGB-D Visual Odometry
ABSTRACT: This paper considers a new bi-objective optimization formulation for robust
RGB-D visual odometry. We investigate two methods for solving the proposed
bi-objective optimization problem: the weighted sum method (in which the
objective functions are combined into a single objective function) and the
bounded objective method (in which one of the objective functions is optimized
and the value of the other objective function is bounded via a constraint). Our
experimental results for the open source TUM RGB-D dataset show that the new
bi-objective optimization formulation is superior to several existing RGB-D
odometry methods. In particular, the new formulation yields more accurate
motion estimates and is more robust when textural or structural features in the
image sequence are lacking.
| no_new_dataset | 0.951188 |
1411.7466 | Chunhua Shen | Lingqiao Liu, Chunhua Shen, Anton van den Hengel | The Treasure beneath Convolutional Layers: Cross-convolutional-layer
Pooling for Image Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A number of recent studies have shown that a Deep Convolutional Neural
Network (DCNN) pretrained on a large dataset can be adopted as a universal
image description which leads to astounding performance in many visual
classification tasks. Most of these studies, if not all, adopt activations of
the fully-connected layer of a DCNN as the image or region representation and
it is believed that convolutional layer activations are less discriminative.
This paper, however, advocates that if used appropriately convolutional layer
activations can be turned into a powerful image representation which enjoys
many advantages over fully-connected layer activations. This is achieved by
adopting a new technique proposed in this paper called
cross-convolutional-layer pooling. More specifically, it extracts subarrays of
feature maps of one convolutional layer as local features and pools the
extracted features with the guidance of feature maps of the successive
convolutional layer. Compared with exising methods that apply DCNNs in the
local feature setting, the proposed method is significantly faster since it
requires much fewer times of DCNN forward computation. Moreover, it avoids the
domain mismatch issue which is usually encountered when applying fully
connected layer activations to describe local regions. By applying our method
to four popular visual classification tasks, it is demonstrated that the
proposed method can achieve comparable or in some cases significantly better
performance than existing fully-connected layer based image representations
while incurring much lower computational cost.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 04:12:57 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Liu",
"Lingqiao",
""
],
[
"Shen",
"Chunhua",
""
],
[
"Hengel",
"Anton van den",
""
]
] | TITLE: The Treasure beneath Convolutional Layers: Cross-convolutional-layer
Pooling for Image Classification
ABSTRACT: A number of recent studies have shown that a Deep Convolutional Neural
Network (DCNN) pretrained on a large dataset can be adopted as a universal
image description which leads to astounding performance in many visual
classification tasks. Most of these studies, if not all, adopt activations of
the fully-connected layer of a DCNN as the image or region representation and
it is believed that convolutional layer activations are less discriminative.
This paper, however, advocates that if used appropriately convolutional layer
activations can be turned into a powerful image representation which enjoys
many advantages over fully-connected layer activations. This is achieved by
adopting a new technique proposed in this paper called
cross-convolutional-layer pooling. More specifically, it extracts subarrays of
feature maps of one convolutional layer as local features and pools the
extracted features with the guidance of feature maps of the successive
convolutional layer. Compared with exising methods that apply DCNNs in the
local feature setting, the proposed method is significantly faster since it
requires much fewer times of DCNN forward computation. Moreover, it avoids the
domain mismatch issue which is usually encountered when applying fully
connected layer activations to describe local regions. By applying our method
to four popular visual classification tasks, it is demonstrated that the
proposed method can achieve comparable or in some cases significantly better
performance than existing fully-connected layer based image representations
while incurring much lower computational cost.
| no_new_dataset | 0.950365 |
1411.7469 | Sanjay Chakraborty | Lopamudra Dey and Sanjay Chakraborty | Canonical PSO Based k-Means Clustering Approach for Real Datasets | null | null | null | null | cs.DB | http://creativecommons.org/licenses/by/3.0/ | "Clustering" the significance and application of this technique is spread
over various fields. Clustering is an unsupervised process in data mining, that
is why the proper evaluation of the results and measuring the compactness and
separability of the clusters are important issues.The procedure of evaluating
the results of a clustering algorithm is known as cluster validity measure.
Different types of indexes are used to solve different types of problems and
indices selection depends on the kind of available data.This paper first
proposes Canonical PSO based K-means clustering algorithm and also analyses
some important clustering indices (intercluster, intracluster) and then
evaluates the effects of those indices on real-time air pollution
database,wholesale customer, wine, and vehicle datasets using typical K-means,
Canonical PSO based K-means, simple PSO based K-means,DBSCAN, and Hierarchical
clustering algorithms.This paper also describes the nature of the clusters and
finally compares the performances of these clustering algorithms according to
the validity assessment. It also defines which algorithm will be more desirable
among all these algorithms to make proper compact clusters on this particular
real life datasets. It actually deals with the behaviour of these clustering
algorithms with respect to validation indexes and represents their results of
evaluation in terms of mathematical and graphical forms.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 04:50:30 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Dey",
"Lopamudra",
""
],
[
"Chakraborty",
"Sanjay",
""
]
] | TITLE: Canonical PSO Based k-Means Clustering Approach for Real Datasets
ABSTRACT: "Clustering" the significance and application of this technique is spread
over various fields. Clustering is an unsupervised process in data mining, that
is why the proper evaluation of the results and measuring the compactness and
separability of the clusters are important issues.The procedure of evaluating
the results of a clustering algorithm is known as cluster validity measure.
Different types of indexes are used to solve different types of problems and
indices selection depends on the kind of available data.This paper first
proposes Canonical PSO based K-means clustering algorithm and also analyses
some important clustering indices (intercluster, intracluster) and then
evaluates the effects of those indices on real-time air pollution
database,wholesale customer, wine, and vehicle datasets using typical K-means,
Canonical PSO based K-means, simple PSO based K-means,DBSCAN, and Hierarchical
clustering algorithms.This paper also describes the nature of the clusters and
finally compares the performances of these clustering algorithms according to
the validity assessment. It also defines which algorithm will be more desirable
among all these algorithms to make proper compact clusters on this particular
real life datasets. It actually deals with the behaviour of these clustering
algorithms with respect to validation indexes and represents their results of
evaluation in terms of mathematical and graphical forms.
| no_new_dataset | 0.952618 |
1411.7474 | Deepinder Kaur Er. | Deepinder Kaur | A Comparative Study of Various Distance Measures for Software fault
prediction | 4 pages,2 figures,"Published with International Journal of Computer
Trends and Technology (IJCTT)" | International Journal of Computer Trends and Technology (IJCTT),
17(3): 117-120, Nov 2014 | 10.14445/22312803/IJCTT-V17P122 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Different distance measures have been used for efficiently predicting
software faults at early stages of software development. One stereotyped
approach for software fault prediction due to its computational efficiency is
K-means clustering, which partitions the dataset into K number of clusters
using any distance measure. Distance measures by using some metrics are used to
extract similar data objects which help in developing efficient algorithms for
clustering and classification. In this paper, we study K-means clustering with
three different distance measures Euclidean, Sorensen and Canberra by using
datasets that have been collected from NASA MDP (metrics data program) .Results
are displayed with the help of ROC curve. The experimental results shows that
K-means clustering with Sorensen distance is better than Euclidean distance and
Canberra distance.
| [
{
"version": "v1",
"created": "Thu, 27 Nov 2014 06:07:46 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Kaur",
"Deepinder",
""
]
] | TITLE: A Comparative Study of Various Distance Measures for Software fault
prediction
ABSTRACT: Different distance measures have been used for efficiently predicting
software faults at early stages of software development. One stereotyped
approach for software fault prediction due to its computational efficiency is
K-means clustering, which partitions the dataset into K number of clusters
using any distance measure. Distance measures by using some metrics are used to
extract similar data objects which help in developing efficient algorithms for
clustering and classification. In this paper, we study K-means clustering with
three different distance measures Euclidean, Sorensen and Canberra by using
datasets that have been collected from NASA MDP (metrics data program) .Results
are displayed with the help of ROC curve. The experimental results shows that
K-means clustering with Sorensen distance is better than Euclidean distance and
Canberra distance.
| no_new_dataset | 0.949763 |
1411.7923 | Dong Yi | Dong Yi, Zhen Lei, Shengcai Liao, and Stan Z. Li | Learning Face Representation from Scratch | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pushing by big data and deep convolutional neural network (CNN), the
performance of face recognition is becoming comparable to human. Using private
large scale training datasets, several groups achieve very high performance on
LFW, i.e., 97% to 99%. While there are many open source implementations of CNN,
none of large scale face dataset is publicly available. The current situation
in the field of face recognition is that data is more important than algorithm.
To solve this problem, this paper proposes a semi-automatical way to collect
face images from Internet and builds a large scale dataset containing about
10,000 subjects and 500,000 images, called CASIAWebFace. Based on the database,
we use a 11-layer CNN to learn discriminative representation and obtain
state-of-theart accuracy on LFW and YTF. The publication of CASIAWebFace will
attract more research groups entering this field and accelerate the development
of face recognition in the wild.
| [
{
"version": "v1",
"created": "Fri, 28 Nov 2014 16:05:18 GMT"
}
] | 2014-12-01T00:00:00 | [
[
"Yi",
"Dong",
""
],
[
"Lei",
"Zhen",
""
],
[
"Liao",
"Shengcai",
""
],
[
"Li",
"Stan Z.",
""
]
] | TITLE: Learning Face Representation from Scratch
ABSTRACT: Pushing by big data and deep convolutional neural network (CNN), the
performance of face recognition is becoming comparable to human. Using private
large scale training datasets, several groups achieve very high performance on
LFW, i.e., 97% to 99%. While there are many open source implementations of CNN,
none of large scale face dataset is publicly available. The current situation
in the field of face recognition is that data is more important than algorithm.
To solve this problem, this paper proposes a semi-automatical way to collect
face images from Internet and builds a large scale dataset containing about
10,000 subjects and 500,000 images, called CASIAWebFace. Based on the database,
we use a 11-layer CNN to learn discriminative representation and obtain
state-of-theart accuracy on LFW and YTF. The publication of CASIAWebFace will
attract more research groups entering this field and accelerate the development
of face recognition in the wild.
| new_dataset | 0.96525 |
1404.4114 | Matthew D. Hoffman | Matthew D. Hoffman and David M. Blei | Structured Stochastic Variational Inference | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Stochastic variational inference makes it possible to approximate posterior
distributions induced by large datasets quickly using stochastic optimization.
The algorithm relies on the use of fully factorized variational distributions.
However, this "mean-field" independence approximation limits the fidelity of
the posterior approximation, and introduces local optima. We show how to relax
the mean-field approximation to allow arbitrary dependencies between global
parameters and local hidden variables, producing better parameter estimates by
reducing bias, sensitivity to local optima, and sensitivity to hyperparameters.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2014 00:12:03 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Nov 2014 18:56:38 GMT"
},
{
"version": "v3",
"created": "Wed, 26 Nov 2014 04:14:16 GMT"
}
] | 2014-11-27T00:00:00 | [
[
"Hoffman",
"Matthew D.",
""
],
[
"Blei",
"David M.",
""
]
] | TITLE: Structured Stochastic Variational Inference
ABSTRACT: Stochastic variational inference makes it possible to approximate posterior
distributions induced by large datasets quickly using stochastic optimization.
The algorithm relies on the use of fully factorized variational distributions.
However, this "mean-field" independence approximation limits the fidelity of
the posterior approximation, and introduces local optima. We show how to relax
the mean-field approximation to allow arbitrary dependencies between global
parameters and local hidden variables, producing better parameter estimates by
reducing bias, sensitivity to local optima, and sensitivity to hyperparameters.
| no_new_dataset | 0.953188 |
1411.7336 | Mohammed alzaidi | Mohammed A. Talab, Siti Norul Huda Sheikh Abdullah, Bilal Bataineh | Edge direction matrixes-based local binar patterns descriptor for shape
pattern recognition | null | null | null | null | cs.CV cs.IR | http://creativecommons.org/licenses/by/3.0/ | Shapes and texture image recognition usage is an essential branch of pattern
recognition. It is made up of techniques that aim at extracting information
from images via human knowledge and works. Local Binary Pattern (LBP) ensures
encoding global and local information and scaling invariance by introducing a
look-up table to reflect the uniformity structure of an object. However, edge
direction matrixes (EDMS) only apply global invariant descriptor which employs
first and secondary order relationships. The main idea behind this methodology
is the need of improved recognition capabilities, a goal achieved by the
combinative use of these descriptors. This collaboration aims to make use of
the major advantages each one presents, by simultaneously complementing each
other, in order to elevate their weak points. By using multiple classifier
approaches such as random forest and multi-layer perceptron neural network, the
proposed combinative descriptor are compared with the state of the art
combinative methods based on Gray-Level Co-occurrence matrix (GLCM with EDMS),
LBP and moment invariant on four benchmark dataset MPEG-7 CE-Shape-1, KTH-TIPS
image, Enghlishfnt and Arabic calligraphy . The experiments have shown the
superiority of the introduced descriptor over the GLCM with EDMS, LBP and
moment invariants and other well-known descriptor such as Scale Invariant
Feature Transform from the literature.
| [
{
"version": "v1",
"created": "Wed, 26 Nov 2014 19:12:33 GMT"
}
] | 2014-11-27T00:00:00 | [
[
"Talab",
"Mohammed A.",
""
],
[
"Abdullah",
"Siti Norul Huda Sheikh",
""
],
[
"Bataineh",
"Bilal",
""
]
] | TITLE: Edge direction matrixes-based local binar patterns descriptor for shape
pattern recognition
ABSTRACT: Shapes and texture image recognition usage is an essential branch of pattern
recognition. It is made up of techniques that aim at extracting information
from images via human knowledge and works. Local Binary Pattern (LBP) ensures
encoding global and local information and scaling invariance by introducing a
look-up table to reflect the uniformity structure of an object. However, edge
direction matrixes (EDMS) only apply global invariant descriptor which employs
first and secondary order relationships. The main idea behind this methodology
is the need of improved recognition capabilities, a goal achieved by the
combinative use of these descriptors. This collaboration aims to make use of
the major advantages each one presents, by simultaneously complementing each
other, in order to elevate their weak points. By using multiple classifier
approaches such as random forest and multi-layer perceptron neural network, the
proposed combinative descriptor are compared with the state of the art
combinative methods based on Gray-Level Co-occurrence matrix (GLCM with EDMS),
LBP and moment invariant on four benchmark dataset MPEG-7 CE-Shape-1, KTH-TIPS
image, Enghlishfnt and Arabic calligraphy . The experiments have shown the
superiority of the introduced descriptor over the GLCM with EDMS, LBP and
moment invariants and other well-known descriptor such as Scale Invariant
Feature Transform from the literature.
| no_new_dataset | 0.949482 |
1406.5549 | Piotr Doll\'ar | Piotr Doll\'ar and C. Lawrence Zitnick | Fast Edge Detection Using Structured Forests | update corresponding to acceptance to PAMI | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Edge detection is a critical component of many vision systems, including
object detectors and image segmentation algorithms. Patches of edges exhibit
well-known forms of local structure, such as straight lines or T-junctions. In
this paper we take advantage of the structure present in local image patches to
learn both an accurate and computationally efficient edge detector. We
formulate the problem of predicting local edge masks in a structured learning
framework applied to random decision forests. Our novel approach to learning
decision trees robustly maps the structured labels to a discrete space on which
standard information gain measures may be evaluated. The result is an approach
that obtains realtime performance that is orders of magnitude faster than many
competing state-of-the-art approaches, while also achieving state-of-the-art
edge detection results on the BSDS500 Segmentation dataset and NYU Depth
dataset. Finally, we show the potential of our approach as a general purpose
edge detector by showing our learned edge models generalize well across
datasets.
| [
{
"version": "v1",
"created": "Fri, 20 Jun 2014 22:28:29 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Nov 2014 02:49:28 GMT"
}
] | 2014-11-26T00:00:00 | [
[
"Dollár",
"Piotr",
""
],
[
"Zitnick",
"C. Lawrence",
""
]
] | TITLE: Fast Edge Detection Using Structured Forests
ABSTRACT: Edge detection is a critical component of many vision systems, including
object detectors and image segmentation algorithms. Patches of edges exhibit
well-known forms of local structure, such as straight lines or T-junctions. In
this paper we take advantage of the structure present in local image patches to
learn both an accurate and computationally efficient edge detector. We
formulate the problem of predicting local edge masks in a structured learning
framework applied to random decision forests. Our novel approach to learning
decision trees robustly maps the structured labels to a discrete space on which
standard information gain measures may be evaluated. The result is an approach
that obtains realtime performance that is orders of magnitude faster than many
competing state-of-the-art approaches, while also achieving state-of-the-art
edge detection results on the BSDS500 Segmentation dataset and NYU Depth
dataset. Finally, we show the potential of our approach as a general purpose
edge detector by showing our learned edge models generalize well across
datasets.
| no_new_dataset | 0.952838 |
1410.7182 | Leon Derczynski | Leon Derczynski, Diana Maynard, Giuseppe Rizzo, Marieke van Erp,
Genevieve Gorrell, Rapha\"el Troncy, Johann Petrak, Kalina Bontcheva | Analysis of Named Entity Recognition and Linking for Tweets | 35 pages, accepted to journal Information Processing and Management | Information Processing & Management 51 (2), 32-49, 2014 | 10.1016/j.ipm.2014.10.006 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Applying natural language processing for mining and intelligent information
access to tweets (a form of microblog) is a challenging, emerging research
area. Unlike carefully authored news text and other longer content, tweets pose
a number of new challenges, due to their short, noisy, context-dependent, and
dynamic nature. Information extraction from tweets is typically performed in a
pipeline, comprising consecutive stages of language identification,
tokenisation, part-of-speech tagging, named entity recognition and entity
disambiguation (e.g. with respect to DBpedia). In this work, we describe a new
Twitter entity disambiguation dataset, and conduct an empirical analysis of
named entity recognition and disambiguation, investigating how robust a number
of state-of-the-art systems are on such noisy texts, what the main sources of
error are, and which problems should be further investigated to improve the
state of the art.
| [
{
"version": "v1",
"created": "Mon, 27 Oct 2014 11:09:36 GMT"
}
] | 2014-11-26T00:00:00 | [
[
"Derczynski",
"Leon",
""
],
[
"Maynard",
"Diana",
""
],
[
"Rizzo",
"Giuseppe",
""
],
[
"van Erp",
"Marieke",
""
],
[
"Gorrell",
"Genevieve",
""
],
[
"Troncy",
"Raphaël",
""
],
[
"Petrak",
"Johann",
""
],
[
"Bontcheva",
"Kalina",
""
]
] | TITLE: Analysis of Named Entity Recognition and Linking for Tweets
ABSTRACT: Applying natural language processing for mining and intelligent information
access to tweets (a form of microblog) is a challenging, emerging research
area. Unlike carefully authored news text and other longer content, tweets pose
a number of new challenges, due to their short, noisy, context-dependent, and
dynamic nature. Information extraction from tweets is typically performed in a
pipeline, comprising consecutive stages of language identification,
tokenisation, part-of-speech tagging, named entity recognition and entity
disambiguation (e.g. with respect to DBpedia). In this work, we describe a new
Twitter entity disambiguation dataset, and conduct an empirical analysis of
named entity recognition and disambiguation, investigating how robust a number
of state-of-the-art systems are on such noisy texts, what the main sources of
error are, and which problems should be further investigated to improve the
state of the art.
| new_dataset | 0.959039 |
1411.6777 | Debajyoti Mukhopadhyay Prof. | Laxmi Lahoti, Chaitali Chandankhede, Debajyoti Mukhopadhyay | Modified Apriori Approach for Evade Network Intrusion Detection System | 5 pages, 3 figures | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Intrusion Detection System or IDS is a software or hardware tool that
repeatedly scans and monitors events that took place in a computer or a
network. A set of rules are used by Signature based Network Intrusion Detection
Systems or NIDS to detect hostile traffic in network segments or packets, which
are so important in detecting malicious and anomalous behaviour over the
network like known attacks that hackers look for new techniques to go unseen.
Sometime, a single failure at any layer will cause the NIDS to miss that
attack. To overcome this problem, a technique is used that will trigger a
failure in that layer. Such technique is known as Evasive technique. An Evasion
can be defined as any technique that modifies a visible attack into any other
form in order to stay away from being detect. The proposed system is used for
detecting attacks which are going on the network and also gives actual
categorization of attacks. The proposed system has advantage of getting low
false alarm rate and high detection rate. So that leads into decrease in
complexity and overhead on the system. The paper presents the Evasion technique
for customized apriori algorithm. The paper aims to make a new functional
structure to evade NIDS. This framework can be used to audit NIDS. This
framework shows that a proof of concept showing how to evade a self built NIDS
considering two publicly available datasets.
| [
{
"version": "v1",
"created": "Tue, 25 Nov 2014 09:16:01 GMT"
}
] | 2014-11-26T00:00:00 | [
[
"Lahoti",
"Laxmi",
""
],
[
"Chandankhede",
"Chaitali",
""
],
[
"Mukhopadhyay",
"Debajyoti",
""
]
] | TITLE: Modified Apriori Approach for Evade Network Intrusion Detection System
ABSTRACT: Intrusion Detection System or IDS is a software or hardware tool that
repeatedly scans and monitors events that took place in a computer or a
network. A set of rules are used by Signature based Network Intrusion Detection
Systems or NIDS to detect hostile traffic in network segments or packets, which
are so important in detecting malicious and anomalous behaviour over the
network like known attacks that hackers look for new techniques to go unseen.
Sometime, a single failure at any layer will cause the NIDS to miss that
attack. To overcome this problem, a technique is used that will trigger a
failure in that layer. Such technique is known as Evasive technique. An Evasion
can be defined as any technique that modifies a visible attack into any other
form in order to stay away from being detect. The proposed system is used for
detecting attacks which are going on the network and also gives actual
categorization of attacks. The proposed system has advantage of getting low
false alarm rate and high detection rate. So that leads into decrease in
complexity and overhead on the system. The paper presents the Evasion technique
for customized apriori algorithm. The paper aims to make a new functional
structure to evade NIDS. This framework can be used to audit NIDS. This
framework shows that a proof of concept showing how to evade a self built NIDS
considering two publicly available datasets.
| no_new_dataset | 0.943243 |
1411.6850 | Khalid Jebari hassani | Amina Dik, Khalid Jebari, Abdelaziz Bouroumi and Aziz Ettouhami | Similarity- based approach for outlier detection | International Journal of Computer Science Issues 2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a new approach for detecting outliers by introducing the
notion of object's proximity. The main idea is that normal point has similar
characteristics with several neighbors. So the point in not an outlier if it
has a high degree of proximity and its neighbors are several. The performance
of this approach is illustrated through real datasets
| [
{
"version": "v1",
"created": "Tue, 25 Nov 2014 13:13:47 GMT"
}
] | 2014-11-26T00:00:00 | [
[
"Dik",
"Amina",
""
],
[
"Jebari",
"Khalid",
""
],
[
"Bouroumi",
"Abdelaziz",
""
],
[
"Ettouhami",
"Aziz",
""
]
] | TITLE: Similarity- based approach for outlier detection
ABSTRACT: This paper presents a new approach for detecting outliers by introducing the
notion of object's proximity. The main idea is that normal point has similar
characteristics with several neighbors. So the point in not an outlier if it
has a high degree of proximity and its neighbors are several. The performance
of this approach is illustrated through real datasets
| no_new_dataset | 0.955319 |
1411.6909 | Aaron Hertzmann | Hamid Izadinia, Ali Farhadi, Aaron Hertzmann, Matthew D. Hoffman | Image Classification and Retrieval from User-Supplied Tags | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes direct learning of image classification from
user-supplied tags, without filtering. Each tag is supplied by the user who
shared the image online. Enormous numbers of these tags are freely available
online, and they give insight about the image categories important to users and
to image classification. Our approach is complementary to the conventional
approach of manual annotation, which is extremely costly. We analyze of the
Flickr 100 Million Image dataset, making several useful observations about the
statistics of these tags. We introduce a large-scale robust classification
algorithm, in order to handle the inherent noise in these tags, and a
calibration procedure to better predict objective annotations. We show that
freely available, user-supplied tags can obtain similar or superior results to
large databases of costly manual annotations.
| [
{
"version": "v1",
"created": "Tue, 25 Nov 2014 16:17:09 GMT"
}
] | 2014-11-26T00:00:00 | [
[
"Izadinia",
"Hamid",
""
],
[
"Farhadi",
"Ali",
""
],
[
"Hertzmann",
"Aaron",
""
],
[
"Hoffman",
"Matthew D.",
""
]
] | TITLE: Image Classification and Retrieval from User-Supplied Tags
ABSTRACT: This paper proposes direct learning of image classification from
user-supplied tags, without filtering. Each tag is supplied by the user who
shared the image online. Enormous numbers of these tags are freely available
online, and they give insight about the image categories important to users and
to image classification. Our approach is complementary to the conventional
approach of manual annotation, which is extremely costly. We analyze of the
Flickr 100 Million Image dataset, making several useful observations about the
statistics of these tags. We introduce a large-scale robust classification
algorithm, in order to handle the inherent noise in these tags, and a
calibration procedure to better predict objective annotations. We show that
freely available, user-supplied tags can obtain similar or superior results to
large databases of costly manual annotations.
| no_new_dataset | 0.953449 |
1210.5268 | Jacob Eisenstein | Jacob Eisenstein, Brendan O'Connor, Noah A. Smith, Eric P. Xing | Diffusion of Lexical Change in Social Media | preprint of PLOS-ONE paper from November 2014; PLoS ONE 9(11) e113114 | null | 10.1371/journal.pone.0113114 | null | cs.CL cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computer-mediated communication is driving fundamental changes in the nature
of written language. We investigate these changes by statistical analysis of a
dataset comprising 107 million Twitter messages (authored by 2.7 million unique
user accounts). Using a latent vector autoregressive model to aggregate across
thousands of words, we identify high-level patterns in diffusion of linguistic
change over the United States. Our model is robust to unpredictable changes in
Twitter's sampling rate, and provides a probabilistic characterization of the
relationship of macro-scale linguistic influence to a set of demographic and
geographic predictors. The results of this analysis offer support for prior
arguments that focus on geographical proximity and population size. However,
demographic similarity -- especially with regard to race -- plays an even more
central role, as cities with similar racial demographics are far more likely to
share linguistic influence. Rather than moving towards a single unified
"netspeak" dialect, language evolution in computer-mediated communication
reproduces existing fault lines in spoken American English.
| [
{
"version": "v1",
"created": "Thu, 18 Oct 2012 21:46:09 GMT"
},
{
"version": "v2",
"created": "Mon, 22 Oct 2012 01:40:54 GMT"
},
{
"version": "v3",
"created": "Tue, 23 Oct 2012 21:15:56 GMT"
},
{
"version": "v4",
"created": "Mon, 24 Nov 2014 03:34:24 GMT"
}
] | 2014-11-25T00:00:00 | [
[
"Eisenstein",
"Jacob",
""
],
[
"O'Connor",
"Brendan",
""
],
[
"Smith",
"Noah A.",
""
],
[
"Xing",
"Eric P.",
""
]
] | TITLE: Diffusion of Lexical Change in Social Media
ABSTRACT: Computer-mediated communication is driving fundamental changes in the nature
of written language. We investigate these changes by statistical analysis of a
dataset comprising 107 million Twitter messages (authored by 2.7 million unique
user accounts). Using a latent vector autoregressive model to aggregate across
thousands of words, we identify high-level patterns in diffusion of linguistic
change over the United States. Our model is robust to unpredictable changes in
Twitter's sampling rate, and provides a probabilistic characterization of the
relationship of macro-scale linguistic influence to a set of demographic and
geographic predictors. The results of this analysis offer support for prior
arguments that focus on geographical proximity and population size. However,
demographic similarity -- especially with regard to race -- plays an even more
central role, as cities with similar racial demographics are far more likely to
share linguistic influence. Rather than moving towards a single unified
"netspeak" dialect, language evolution in computer-mediated communication
reproduces existing fault lines in spoken American English.
| no_new_dataset | 0.874185 |
1411.6235 | Xiaojun Chang | Xiaojun Chang, Feiping Nie, Zhigang Ma, and Yi Yang | Balanced k-Means and Min-Cut Clustering | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Clustering is an effective technique in data mining to generate groups that
are the matter of interest. Among various clustering approaches, the family of
k-means algorithms and min-cut algorithms gain most popularity due to their
simplicity and efficacy. The classical k-means algorithm partitions a number of
data points into several subsets by iteratively updating the clustering centers
and the associated data points. By contrast, a weighted undirected graph is
constructed in min-cut algorithms which partition the vertices of the graph
into two sets. However, existing clustering algorithms tend to cluster minority
of data points into a subset, which shall be avoided when the target dataset is
balanced. To achieve more accurate clustering for balanced dataset, we propose
to leverage exclusive lasso on k-means and min-cut to regulate the balance
degree of the clustering results. By optimizing our objective functions that
build atop the exclusive lasso, we can make the clustering result as much
balanced as possible. Extensive experiments on several large-scale datasets
validate the advantage of the proposed algorithms compared to the
state-of-the-art clustering algorithms.
| [
{
"version": "v1",
"created": "Sun, 23 Nov 2014 13:16:25 GMT"
}
] | 2014-11-25T00:00:00 | [
[
"Chang",
"Xiaojun",
""
],
[
"Nie",
"Feiping",
""
],
[
"Ma",
"Zhigang",
""
],
[
"Yang",
"Yi",
""
]
] | TITLE: Balanced k-Means and Min-Cut Clustering
ABSTRACT: Clustering is an effective technique in data mining to generate groups that
are the matter of interest. Among various clustering approaches, the family of
k-means algorithms and min-cut algorithms gain most popularity due to their
simplicity and efficacy. The classical k-means algorithm partitions a number of
data points into several subsets by iteratively updating the clustering centers
and the associated data points. By contrast, a weighted undirected graph is
constructed in min-cut algorithms which partition the vertices of the graph
into two sets. However, existing clustering algorithms tend to cluster minority
of data points into a subset, which shall be avoided when the target dataset is
balanced. To achieve more accurate clustering for balanced dataset, we propose
to leverage exclusive lasso on k-means and min-cut to regulate the balance
degree of the clustering results. By optimizing our objective functions that
build atop the exclusive lasso, we can make the clustering result as much
balanced as possible. Extensive experiments on several large-scale datasets
validate the advantage of the proposed algorithms compared to the
state-of-the-art clustering algorithms.
| no_new_dataset | 0.951414 |
1411.6308 | Xiaojun Chang | Xiaojun Chang, Feiping Nie, Zhigang Ma, Yi Yang and Xiaofang Zhou | A Convex Formulation for Spectral Shrunk Clustering | AAAI2015 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spectral clustering is a fundamental technique in the field of data mining
and information processing. Most existing spectral clustering algorithms
integrate dimensionality reduction into the clustering process assisted by
manifold learning in the original space. However, the manifold in
reduced-dimensional subspace is likely to exhibit altered properties in
contrast with the original space. Thus, applying manifold information obtained
from the original space to the clustering process in a low-dimensional subspace
is prone to inferior performance. Aiming to address this issue, we propose a
novel convex algorithm that mines the manifold structure in the low-dimensional
subspace. In addition, our unified learning process makes the manifold learning
particularly tailored for the clustering. Compared with other related methods,
the proposed algorithm results in more structured clustering result. To
validate the efficacy of the proposed algorithm, we perform extensive
experiments on several benchmark datasets in comparison with some
state-of-the-art clustering approaches. The experimental results demonstrate
that the proposed algorithm has quite promising clustering performance.
| [
{
"version": "v1",
"created": "Sun, 23 Nov 2014 22:12:52 GMT"
}
] | 2014-11-25T00:00:00 | [
[
"Chang",
"Xiaojun",
""
],
[
"Nie",
"Feiping",
""
],
[
"Ma",
"Zhigang",
""
],
[
"Yang",
"Yi",
""
],
[
"Zhou",
"Xiaofang",
""
]
] | TITLE: A Convex Formulation for Spectral Shrunk Clustering
ABSTRACT: Spectral clustering is a fundamental technique in the field of data mining
and information processing. Most existing spectral clustering algorithms
integrate dimensionality reduction into the clustering process assisted by
manifold learning in the original space. However, the manifold in
reduced-dimensional subspace is likely to exhibit altered properties in
contrast with the original space. Thus, applying manifold information obtained
from the original space to the clustering process in a low-dimensional subspace
is prone to inferior performance. Aiming to address this issue, we propose a
novel convex algorithm that mines the manifold structure in the low-dimensional
subspace. In addition, our unified learning process makes the manifold learning
particularly tailored for the clustering. Compared with other related methods,
the proposed algorithm results in more structured clustering result. To
validate the efficacy of the proposed algorithm, we perform extensive
experiments on several benchmark datasets in comparison with some
state-of-the-art clustering approaches. The experimental results demonstrate
that the proposed algorithm has quite promising clustering performance.
| no_new_dataset | 0.952838 |
1411.6447 | Tianjun Xiao | Tianjun Xiao, Yichong Xu, Kuiyuan Yang, Jiaxing Zhang, Yuxin Peng,
Zheng Zhang | The Application of Two-level Attention Models in Deep Convolutional
Neural Network for Fine-grained Image Classification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fine-grained classification is challenging because categories can only be
discriminated by subtle and local differences. Variances in the pose, scale or
rotation usually make the problem more difficult. Most fine-grained
classification systems follow the pipeline of finding foreground object or
object parts (where) to extract discriminative features (what).
In this paper, we propose to apply visual attention to fine-grained
classification task using deep neural network. Our pipeline integrates three
types of attention: the bottom-up attention that propose candidate patches, the
object-level top-down attention that selects relevant patches to a certain
object, and the part-level top-down attention that localizes discriminative
parts. We combine these attentions to train domain-specific deep nets, then use
it to improve both the what and where aspects. Importantly, we avoid using
expensive annotations like bounding box or part information from end-to-end.
The weak supervision constraint makes our work easier to generalize.
We have verified the effectiveness of the method on the subsets of ILSVRC2012
dataset and CUB200_2011 dataset. Our pipeline delivered significant
improvements and achieved the best accuracy under the weakest supervision
condition. The performance is competitive against other methods that rely on
additional annotations.
| [
{
"version": "v1",
"created": "Mon, 24 Nov 2014 13:30:07 GMT"
}
] | 2014-11-25T00:00:00 | [
[
"Xiao",
"Tianjun",
""
],
[
"Xu",
"Yichong",
""
],
[
"Yang",
"Kuiyuan",
""
],
[
"Zhang",
"Jiaxing",
""
],
[
"Peng",
"Yuxin",
""
],
[
"Zhang",
"Zheng",
""
]
] | TITLE: The Application of Two-level Attention Models in Deep Convolutional
Neural Network for Fine-grained Image Classification
ABSTRACT: Fine-grained classification is challenging because categories can only be
discriminated by subtle and local differences. Variances in the pose, scale or
rotation usually make the problem more difficult. Most fine-grained
classification systems follow the pipeline of finding foreground object or
object parts (where) to extract discriminative features (what).
In this paper, we propose to apply visual attention to fine-grained
classification task using deep neural network. Our pipeline integrates three
types of attention: the bottom-up attention that propose candidate patches, the
object-level top-down attention that selects relevant patches to a certain
object, and the part-level top-down attention that localizes discriminative
parts. We combine these attentions to train domain-specific deep nets, then use
it to improve both the what and where aspects. Importantly, we avoid using
expensive annotations like bounding box or part information from end-to-end.
The weak supervision constraint makes our work easier to generalize.
We have verified the effectiveness of the method on the subsets of ILSVRC2012
dataset and CUB200_2011 dataset. Our pipeline delivered significant
improvements and achieved the best accuracy under the weakest supervision
condition. The performance is competitive against other methods that rely on
additional annotations.
| no_new_dataset | 0.948394 |
1411.6562 | Manas Joglekar | Manas Joglekar, Hector Garcia-Molina, Aditya Parameswaran | Evaluating the Crowd with Confidence | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Worker quality control is a crucial aspect of crowdsourcing systems;
typically occupying a large fraction of the time and money invested on
crowdsourcing. In this work, we devise techniques to generate confidence
intervals for worker error rate estimates, thereby enabling a better evaluation
of worker quality. We show that our techniques generate correct confidence
intervals on a range of real-world datasets, and demonstrate wide applicability
by using them to evict poorly performing workers, and provide confidence
intervals on the accuracy of the answers.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 23:50:43 GMT"
}
] | 2014-11-25T00:00:00 | [
[
"Joglekar",
"Manas",
""
],
[
"Garcia-Molina",
"Hector",
""
],
[
"Parameswaran",
"Aditya",
""
]
] | TITLE: Evaluating the Crowd with Confidence
ABSTRACT: Worker quality control is a crucial aspect of crowdsourcing systems;
typically occupying a large fraction of the time and money invested on
crowdsourcing. In this work, we devise techniques to generate confidence
intervals for worker error rate estimates, thereby enabling a better evaluation
of worker quality. We show that our techniques generate correct confidence
intervals on a range of real-world datasets, and demonstrate wide applicability
by using them to evict poorly performing workers, and provide confidence
intervals on the accuracy of the answers.
| no_new_dataset | 0.959116 |
1405.0538 | Rados{\l}aw Michalski | Rados{\l}aw Michalski, Tomasz Kajdanowicz, Piotr Br\'odka,
Przemys{\l}aw Kazienko | Seed Selection for Spread of Influence in Social Networks: Temporal vs.
Static Approach | null | New Generation Computing, Vol. 32, Issue 3-4, pp. 213-235, 2014 | 10.1007/s00354-014-0402-9 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of finding optimal set of users for influencing others in the
social network has been widely studied. Because it is NP-hard, some heuristics
were proposed to find sub-optimal solutions. Still, one of the commonly used
assumption is the one that seeds are chosen on the static network, not the
dynamic one. This static approach is in fact far from the real-world networks,
where new nodes may appear and old ones dynamically disappear in course of
time.
The main purpose of this paper is to analyse how the results of one of the
typical models for spread of influence - linear threshold - differ depending on
the strategy of building the social network used later for choosing seeds. To
show the impact of network creation strategy on the final number of influenced
nodes - outcome of spread of influence, the results for three approaches were
studied: one static and two temporal with different granularities, i.e. various
number of time windows. Social networks for each time window encapsulated
dynamic changes in the network structure. Calculation of various node
structural measures like degree or betweenness respected these changes by means
of forgetting mechanism - more recent data had greater influence on node
measure values. These measures were, in turn, used for node ranking and their
selection for seeding.
All concepts were applied to experimental verification on five real datasets.
The results revealed that temporal approach is always better than static and
the higher granularity in the temporal social network while seeding, the more
finally influenced nodes. Additionally, outdegree measure with exponential
forgetting typically outperformed other time-dependent structural measures, if
used for seed candidate ranking.
| [
{
"version": "v1",
"created": "Fri, 2 May 2014 23:32:04 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Nov 2014 13:20:36 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Michalski",
"Radosław",
""
],
[
"Kajdanowicz",
"Tomasz",
""
],
[
"Bródka",
"Piotr",
""
],
[
"Kazienko",
"Przemysław",
""
]
] | TITLE: Seed Selection for Spread of Influence in Social Networks: Temporal vs.
Static Approach
ABSTRACT: The problem of finding optimal set of users for influencing others in the
social network has been widely studied. Because it is NP-hard, some heuristics
were proposed to find sub-optimal solutions. Still, one of the commonly used
assumption is the one that seeds are chosen on the static network, not the
dynamic one. This static approach is in fact far from the real-world networks,
where new nodes may appear and old ones dynamically disappear in course of
time.
The main purpose of this paper is to analyse how the results of one of the
typical models for spread of influence - linear threshold - differ depending on
the strategy of building the social network used later for choosing seeds. To
show the impact of network creation strategy on the final number of influenced
nodes - outcome of spread of influence, the results for three approaches were
studied: one static and two temporal with different granularities, i.e. various
number of time windows. Social networks for each time window encapsulated
dynamic changes in the network structure. Calculation of various node
structural measures like degree or betweenness respected these changes by means
of forgetting mechanism - more recent data had greater influence on node
measure values. These measures were, in turn, used for node ranking and their
selection for seeding.
All concepts were applied to experimental verification on five real datasets.
The results revealed that temporal approach is always better than static and
the higher granularity in the temporal social network while seeding, the more
finally influenced nodes. Additionally, outdegree measure with exponential
forgetting typically outperformed other time-dependent structural measures, if
used for seed candidate ranking.
| no_new_dataset | 0.958148 |
1410.4510 | Finale Doshi-Velez | Finale Doshi-Velez and Byron Wallace and Ryan Adams | Graph-Sparse LDA: A Topic Model with Structured Sparsity | null | null | null | null | stat.ML cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Originally designed to model text, topic modeling has become a powerful tool
for uncovering latent structure in domains including medicine, finance, and
vision. The goals for the model vary depending on the application: in some
cases, the discovered topics may be used for prediction or some other
downstream task. In other cases, the content of the topic itself may be of
intrinsic scientific interest.
Unfortunately, even using modern sparse techniques, the discovered topics are
often difficult to interpret due to the high dimensionality of the underlying
space. To improve topic interpretability, we introduce Graph-Sparse LDA, a
hierarchical topic model that leverages knowledge of relationships between
words (e.g., as encoded by an ontology). In our model, topics are summarized by
a few latent concept-words from the underlying graph that explain the observed
words. Graph-Sparse LDA recovers sparse, interpretable summaries on two
real-world biomedical datasets while matching state-of-the-art prediction
performance.
| [
{
"version": "v1",
"created": "Thu, 16 Oct 2014 17:35:31 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Nov 2014 16:38:59 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Doshi-Velez",
"Finale",
""
],
[
"Wallace",
"Byron",
""
],
[
"Adams",
"Ryan",
""
]
] | TITLE: Graph-Sparse LDA: A Topic Model with Structured Sparsity
ABSTRACT: Originally designed to model text, topic modeling has become a powerful tool
for uncovering latent structure in domains including medicine, finance, and
vision. The goals for the model vary depending on the application: in some
cases, the discovered topics may be used for prediction or some other
downstream task. In other cases, the content of the topic itself may be of
intrinsic scientific interest.
Unfortunately, even using modern sparse techniques, the discovered topics are
often difficult to interpret due to the high dimensionality of the underlying
space. To improve topic interpretability, we introduce Graph-Sparse LDA, a
hierarchical topic model that leverages knowledge of relationships between
words (e.g., as encoded by an ontology). In our model, topics are summarized by
a few latent concept-words from the underlying graph that explain the observed
words. Graph-Sparse LDA recovers sparse, interpretable summaries on two
real-world biomedical datasets while matching state-of-the-art prediction
performance.
| no_new_dataset | 0.952086 |
1411.5428 | Ben Stoddard | Ben Stoddard and Yan Chen and Ashwin Machanavajjhala | Differentially Private Algorithms for Empirical Machine Learning | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An important use of private data is to build machine learning classifiers.
While there is a burgeoning literature on differentially private classification
algorithms, we find that they are not practical in real applications due to two
reasons. First, existing differentially private classifiers provide poor
accuracy on real world datasets. Second, there is no known differentially
private algorithm for empirically evaluating the private classifier on a
private test dataset.
In this paper, we develop differentially private algorithms that mirror real
world empirical machine learning workflows. We consider the private classifier
training algorithm as a blackbox. We present private algorithms for selecting
features that are input to the classifier. Though adding a preprocessing step
takes away some of the privacy budget from the actual classification process
(thus potentially making it noisier and less accurate), we show that our novel
preprocessing techniques significantly increase classifier accuracy on three
real-world datasets. We also present the first private algorithms for
empirically constructing receiver operating characteristic (ROC) curves on a
private test set.
| [
{
"version": "v1",
"created": "Thu, 20 Nov 2014 03:10:47 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Nov 2014 20:41:04 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Stoddard",
"Ben",
""
],
[
"Chen",
"Yan",
""
],
[
"Machanavajjhala",
"Ashwin",
""
]
] | TITLE: Differentially Private Algorithms for Empirical Machine Learning
ABSTRACT: An important use of private data is to build machine learning classifiers.
While there is a burgeoning literature on differentially private classification
algorithms, we find that they are not practical in real applications due to two
reasons. First, existing differentially private classifiers provide poor
accuracy on real world datasets. Second, there is no known differentially
private algorithm for empirically evaluating the private classifier on a
private test dataset.
In this paper, we develop differentially private algorithms that mirror real
world empirical machine learning workflows. We consider the private classifier
training algorithm as a blackbox. We present private algorithms for selecting
features that are input to the classifier. Though adding a preprocessing step
takes away some of the privacy budget from the actual classification process
(thus potentially making it noisier and less accurate), we show that our novel
preprocessing techniques significantly increase classifier accuracy on three
real-world datasets. We also present the first private algorithms for
empirically constructing receiver operating characteristic (ROC) curves on a
private test set.
| no_new_dataset | 0.948394 |
1411.5731 | Suleyman Cetintas | Can Xu, Suleyman Cetintas, Kuang-Chih Lee, Li-Jia Li | Visual Sentiment Prediction with Deep Convolutional Neural Networks | null | null | null | null | cs.CV cs.NE stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Images have become one of the most popular types of media through which users
convey their emotions within online social networks. Although vast amount of
research is devoted to sentiment analysis of textual data, there has been very
limited work that focuses on analyzing sentiment of image data. In this work,
we propose a novel visual sentiment prediction framework that performs image
understanding with Deep Convolutional Neural Networks (CNN). Specifically, the
proposed sentiment prediction framework performs transfer learning from a CNN
with millions of parameters, which is pre-trained on large-scale data for
object recognition. Experiments conducted on two real-world datasets from
Twitter and Tumblr demonstrate the effectiveness of the proposed visual
sentiment analysis framework.
| [
{
"version": "v1",
"created": "Fri, 21 Nov 2014 00:39:43 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Xu",
"Can",
""
],
[
"Cetintas",
"Suleyman",
""
],
[
"Lee",
"Kuang-Chih",
""
],
[
"Li",
"Li-Jia",
""
]
] | TITLE: Visual Sentiment Prediction with Deep Convolutional Neural Networks
ABSTRACT: Images have become one of the most popular types of media through which users
convey their emotions within online social networks. Although vast amount of
research is devoted to sentiment analysis of textual data, there has been very
limited work that focuses on analyzing sentiment of image data. In this work,
we propose a novel visual sentiment prediction framework that performs image
understanding with Deep Convolutional Neural Networks (CNN). Specifically, the
proposed sentiment prediction framework performs transfer learning from a CNN
with millions of parameters, which is pre-trained on large-scale data for
object recognition. Experiments conducted on two real-world datasets from
Twitter and Tumblr demonstrate the effectiveness of the proposed visual
sentiment analysis framework.
| no_new_dataset | 0.952309 |
1411.5935 | M. Zeeshan Zia | M.Zeeshan Zia, Michael Stark, Konrad Schindler | Towards Scene Understanding with Detailed 3D Object Representations | International Journal of Computer Vision (appeared online on 4
November 2014). Online version:
http://link.springer.com/article/10.1007/s11263-014-0780-y | null | 10.1007/s11263-014-0780-y | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current approaches to semantic image and scene understanding typically employ
rather simple object representations such as 2D or 3D bounding boxes. While
such coarse models are robust and allow for reliable object detection, they
discard much of the information about objects' 3D shape and pose, and thus do
not lend themselves well to higher-level reasoning. Here, we propose to base
scene understanding on a high-resolution object representation. An object class
- in our case cars - is modeled as a deformable 3D wireframe, which enables
fine-grained modeling at the level of individual vertices and faces. We augment
that model to explicitly include vertex-level occlusion, and embed all
instances in a common coordinate frame, in order to infer and exploit
object-object interactions. Specifically, from a single view we jointly
estimate the shapes and poses of multiple objects in a common 3D frame. A
ground plane in that frame is estimated by consensus among different objects,
which significantly stabilizes monocular 3D pose estimation. The fine-grained
model, in conjunction with the explicit 3D scene model, further allows one to
infer part-level occlusions between the modeled objects, as well as occlusions
by other, unmodeled scene elements. To demonstrate the benefits of such
detailed object class models in the context of scene understanding we
systematically evaluate our approach on the challenging KITTI street scene
dataset. The experiments show that the model's ability to utilize image
evidence at the level of individual parts improves monocular 3D pose estimation
w.r.t. both location and (continuous) viewpoint.
| [
{
"version": "v1",
"created": "Tue, 18 Nov 2014 15:07:19 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Zia",
"M. Zeeshan",
""
],
[
"Stark",
"Michael",
""
],
[
"Schindler",
"Konrad",
""
]
] | TITLE: Towards Scene Understanding with Detailed 3D Object Representations
ABSTRACT: Current approaches to semantic image and scene understanding typically employ
rather simple object representations such as 2D or 3D bounding boxes. While
such coarse models are robust and allow for reliable object detection, they
discard much of the information about objects' 3D shape and pose, and thus do
not lend themselves well to higher-level reasoning. Here, we propose to base
scene understanding on a high-resolution object representation. An object class
- in our case cars - is modeled as a deformable 3D wireframe, which enables
fine-grained modeling at the level of individual vertices and faces. We augment
that model to explicitly include vertex-level occlusion, and embed all
instances in a common coordinate frame, in order to infer and exploit
object-object interactions. Specifically, from a single view we jointly
estimate the shapes and poses of multiple objects in a common 3D frame. A
ground plane in that frame is estimated by consensus among different objects,
which significantly stabilizes monocular 3D pose estimation. The fine-grained
model, in conjunction with the explicit 3D scene model, further allows one to
infer part-level occlusions between the modeled objects, as well as occlusions
by other, unmodeled scene elements. To demonstrate the benefits of such
detailed object class models in the context of scene understanding we
systematically evaluate our approach on the challenging KITTI street scene
dataset. The experiments show that the model's ability to utilize image
evidence at the level of individual parts improves monocular 3D pose estimation
w.r.t. both location and (continuous) viewpoint.
| no_new_dataset | 0.944536 |
1411.5995 | George Ovchinnikov | G.V. Ovchinnikov, D.A. Kolesnikov, I.V. Oseledets | Algebraic reputation model RepRank and its application to spambot
detection | null | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to popularity surge social networks became lucrative targets for spammers
and guerilla marketers, who are trying to game ranking systems and broadcast
their messages at little to none cost. Ranking systems, for example Twitter's
Trends, can be gamed by scripted users also called bots, who are automatically
or semi-automatically twitting essentially the same message. Judging by the
prices and abundance of supply from PR firms this is an easy to implement and
widely used tactic, at least in Russian blogosphere. Aggregative analysis of
social networks should at best mark those messages as spam or at least
correctly downplay their importance as they represent opinions only of a few,
if dedicated, users. Hence bot detection plays a crucial role in social network
mining and analysis. In this paper we propose technique called RepRank which
could be viewed as Markov chain based model for reputation propagation on
graphs utilizing simultaneous trust and anti-trust propagation and provide
effective numerical approach for its computation. Comparison with another
models such as TrustRank and some of its modifications on sample of 320000
Russian speaking Twitter users is presented. The dataset is presented as well.
| [
{
"version": "v1",
"created": "Thu, 20 Nov 2014 13:50:39 GMT"
}
] | 2014-11-24T00:00:00 | [
[
"Ovchinnikov",
"G. V.",
""
],
[
"Kolesnikov",
"D. A.",
""
],
[
"Oseledets",
"I. V.",
""
]
] | TITLE: Algebraic reputation model RepRank and its application to spambot
detection
ABSTRACT: Due to popularity surge social networks became lucrative targets for spammers
and guerilla marketers, who are trying to game ranking systems and broadcast
their messages at little to none cost. Ranking systems, for example Twitter's
Trends, can be gamed by scripted users also called bots, who are automatically
or semi-automatically twitting essentially the same message. Judging by the
prices and abundance of supply from PR firms this is an easy to implement and
widely used tactic, at least in Russian blogosphere. Aggregative analysis of
social networks should at best mark those messages as spam or at least
correctly downplay their importance as they represent opinions only of a few,
if dedicated, users. Hence bot detection plays a crucial role in social network
mining and analysis. In this paper we propose technique called RepRank which
could be viewed as Markov chain based model for reputation propagation on
graphs utilizing simultaneous trust and anti-trust propagation and provide
effective numerical approach for its computation. Comparison with another
models such as TrustRank and some of its modifications on sample of 320000
Russian speaking Twitter users is presented. The dataset is presented as well.
| new_dataset | 0.666049 |
1307.3176 | L.A. Prashanth | Nathaniel Korda, Prashanth L.A. and R\'emi Munos | Fast gradient descent for drifting least squares regression, with
application to bandits | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Online learning algorithms require to often recompute least squares
regression estimates of parameters. We study improving the computational
complexity of such algorithms by using stochastic gradient descent (SGD) type
schemes in place of classic regression solvers. We show that SGD schemes
efficiently track the true solutions of the regression problems, even in the
presence of a drift. This finding coupled with an $O(d)$ improvement in
complexity, where $d$ is the dimension of the data, make them attractive for
implementation in the big data settings. In the case when strong convexity in
the regression problem is guaranteed, we provide bounds on the error both in
expectation and high probability (the latter is often needed to provide
theoretical guarantees for higher level algorithms), despite the drifting least
squares solution. As an example of this case we prove that the regret
performance of an SGD version of the PEGE linear bandit algorithm
[Rusmevichientong and Tsitsiklis 2010] is worse that that of PEGE itself only
by a factor of $O(\log^4 n)$. When strong convexity of the regression problem
cannot be guaranteed, we investigate using an adaptive regularisation. We make
an empirical study of an adaptively regularised, SGD version of LinUCB [Li et
al. 2010] in a news article recommendation application, which uses the large
scale news recommendation dataset from Yahoo! front page. These experiments
show a large gain in computational complexity, with a consistently low tracking
error and click-through-rate (CTR) performance that is $75\%$ close.
| [
{
"version": "v1",
"created": "Thu, 11 Jul 2013 16:36:29 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Feb 2014 00:27:18 GMT"
},
{
"version": "v3",
"created": "Thu, 24 Jul 2014 14:29:52 GMT"
},
{
"version": "v4",
"created": "Thu, 20 Nov 2014 12:40:48 GMT"
}
] | 2014-11-21T00:00:00 | [
[
"Korda",
"Nathaniel",
""
],
[
"A.",
"Prashanth L.",
""
],
[
"Munos",
"Rémi",
""
]
] | TITLE: Fast gradient descent for drifting least squares regression, with
application to bandits
ABSTRACT: Online learning algorithms require to often recompute least squares
regression estimates of parameters. We study improving the computational
complexity of such algorithms by using stochastic gradient descent (SGD) type
schemes in place of classic regression solvers. We show that SGD schemes
efficiently track the true solutions of the regression problems, even in the
presence of a drift. This finding coupled with an $O(d)$ improvement in
complexity, where $d$ is the dimension of the data, make them attractive for
implementation in the big data settings. In the case when strong convexity in
the regression problem is guaranteed, we provide bounds on the error both in
expectation and high probability (the latter is often needed to provide
theoretical guarantees for higher level algorithms), despite the drifting least
squares solution. As an example of this case we prove that the regret
performance of an SGD version of the PEGE linear bandit algorithm
[Rusmevichientong and Tsitsiklis 2010] is worse that that of PEGE itself only
by a factor of $O(\log^4 n)$. When strong convexity of the regression problem
cannot be guaranteed, we investigate using an adaptive regularisation. We make
an empirical study of an adaptively regularised, SGD version of LinUCB [Li et
al. 2010] in a news article recommendation application, which uses the large
scale news recommendation dataset from Yahoo! front page. These experiments
show a large gain in computational complexity, with a consistently low tracking
error and click-through-rate (CTR) performance that is $75\%$ close.
| no_new_dataset | 0.937783 |
1004.3499 | Lee Samuel Finn | Lee Samuel Finn, Andrea N. Lommen | Detection, Localization and Characterization of Gravitational Wave
Bursts in a Pulsar Timing Array | 43 pages, 13 figures, submitted to ApJ. | Astrophys.J.718:1400-1415,2010 | 10.1088/0004-637X/718/2/1400 | null | astro-ph.IM gr-qc physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Efforts to detect gravitational waves by timing an array of pulsars have
focused traditionally on stationary gravitational waves: e.g., stochastic or
periodic signals. Gravitational wave bursts --- signals whose duration is much
shorter than the observation period --- will also arise in the pulsar timing
array waveband. Sources that give rise to detectable bursts include the
formation or coalescence of supermassive black holes (SMBHs), the periapsis
passage of compact objects in highly elliptic or unbound orbits about a SMBH,
or cusps on cosmic strings. Here we describe how pulsar timing array data may
be analyzed to detect and characterize these bursts. Our analysis addresses, in
a mutually consistent manner, a hierarchy of three questions: \emph{i}) What
are the odds that a dataset includes the signal from a gravitational wave
burst? \emph{ii}) Assuming the presence of a burst, what is the direction to
its source? and \emph{iii}) Assuming the burst propagation direction, what is
the burst waveform's time dependence in each of its polarization states?
Applying our analysis to synthetic data sets we find that we can \emph{detect}
gravitational waves even when the radiation is too weak to either localize the
source of infer the waveform, and \emph{detect} and \emph{localize} sources
even when the radiation amplitude is too weak to permit the waveform to be
determined. While the context of our discussion is gravitational wave detection
via pulsar timing arrays, the analysis itself is directly applicable to
gravitational wave detection using either ground or space-based detector data.
| [
{
"version": "v1",
"created": "Tue, 20 Apr 2010 16:40:35 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Finn",
"Lee Samuel",
""
],
[
"Lommen",
"Andrea N.",
""
]
] | TITLE: Detection, Localization and Characterization of Gravitational Wave
Bursts in a Pulsar Timing Array
ABSTRACT: Efforts to detect gravitational waves by timing an array of pulsars have
focused traditionally on stationary gravitational waves: e.g., stochastic or
periodic signals. Gravitational wave bursts --- signals whose duration is much
shorter than the observation period --- will also arise in the pulsar timing
array waveband. Sources that give rise to detectable bursts include the
formation or coalescence of supermassive black holes (SMBHs), the periapsis
passage of compact objects in highly elliptic or unbound orbits about a SMBH,
or cusps on cosmic strings. Here we describe how pulsar timing array data may
be analyzed to detect and characterize these bursts. Our analysis addresses, in
a mutually consistent manner, a hierarchy of three questions: \emph{i}) What
are the odds that a dataset includes the signal from a gravitational wave
burst? \emph{ii}) Assuming the presence of a burst, what is the direction to
its source? and \emph{iii}) Assuming the burst propagation direction, what is
the burst waveform's time dependence in each of its polarization states?
Applying our analysis to synthetic data sets we find that we can \emph{detect}
gravitational waves even when the radiation is too weak to either localize the
source of infer the waveform, and \emph{detect} and \emph{localize} sources
even when the radiation amplitude is too weak to permit the waveform to be
determined. While the context of our discussion is gravitational wave detection
via pulsar timing arrays, the analysis itself is directly applicable to
gravitational wave detection using either ground or space-based detector data.
| no_new_dataset | 0.94743 |
1406.6312 | Ahmed El-Kishky | Ahmed El-Kishky, Yanglei Song, Chi Wang, Clare Voss, Jiawei Han | Scalable Topical Phrase Mining from Text Corpora | null | Proceedings of the VLDB Endowment, Vol. 8(3), pp. 305 - 316, 2014 | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While most topic modeling algorithms model text corpora with unigrams, human
interpretation often relies on inherent grouping of terms into phrases. As
such, we consider the problem of discovering topical phrases of mixed lengths.
Existing work either performs post processing to the inference results of
unigram-based topic models, or utilizes complex n-gram-discovery topic models.
These methods generally produce low-quality topical phrases or suffer from poor
scalability on even moderately-sized datasets. We propose a different approach
that is both computationally efficient and effective. Our solution combines a
novel phrase mining framework to segment a document into single and multi-word
phrases, and a new topic model that operates on the induced document partition.
Our approach discovers high quality topical phrases with negligible extra cost
to the bag-of-words topic model in a variety of datasets including research
publication titles, abstracts, reviews, and news articles.
| [
{
"version": "v1",
"created": "Tue, 24 Jun 2014 17:10:29 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Nov 2014 00:18:06 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"El-Kishky",
"Ahmed",
""
],
[
"Song",
"Yanglei",
""
],
[
"Wang",
"Chi",
""
],
[
"Voss",
"Clare",
""
],
[
"Han",
"Jiawei",
""
]
] | TITLE: Scalable Topical Phrase Mining from Text Corpora
ABSTRACT: While most topic modeling algorithms model text corpora with unigrams, human
interpretation often relies on inherent grouping of terms into phrases. As
such, we consider the problem of discovering topical phrases of mixed lengths.
Existing work either performs post processing to the inference results of
unigram-based topic models, or utilizes complex n-gram-discovery topic models.
These methods generally produce low-quality topical phrases or suffer from poor
scalability on even moderately-sized datasets. We propose a different approach
that is both computationally efficient and effective. Our solution combines a
novel phrase mining framework to segment a document into single and multi-word
phrases, and a new topic model that operates on the induced document partition.
Our approach discovers high quality topical phrases with negligible extra cost
to the bag-of-words topic model in a variety of datasets including research
publication titles, abstracts, reviews, and news articles.
| no_new_dataset | 0.950549 |
1407.7094 | Bruno Gon\c{c}alves | Bruno Gon\c{c}alves and David S\'anchez | Crowdsourcing Dialect Characterization through Twitter | 10 pages, 5 figures | PLoS One 9, E112074 (2014) | 10.1371/journal.pone.0112074 | null | physics.soc-ph cs.CL cs.SI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We perform a large-scale analysis of language diatopic variation using
geotagged microblogging datasets. By collecting all Twitter messages written in
Spanish over more than two years, we build a corpus from which a carefully
selected list of concepts allows us to characterize Spanish varieties on a
global scale. A cluster analysis proves the existence of well defined
macroregions sharing common lexical properties. Remarkably enough, we find that
Spanish language is split into two superdialects, namely, an urban speech used
across major American and Spanish citites and a diverse form that encompasses
rural areas and small towns. The latter can be further clustered into smaller
varieties with a stronger regional character.
| [
{
"version": "v1",
"created": "Sat, 26 Jul 2014 04:16:31 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Gonçalves",
"Bruno",
""
],
[
"Sánchez",
"David",
""
]
] | TITLE: Crowdsourcing Dialect Characterization through Twitter
ABSTRACT: We perform a large-scale analysis of language diatopic variation using
geotagged microblogging datasets. By collecting all Twitter messages written in
Spanish over more than two years, we build a corpus from which a carefully
selected list of concepts allows us to characterize Spanish varieties on a
global scale. A cluster analysis proves the existence of well defined
macroregions sharing common lexical properties. Remarkably enough, we find that
Spanish language is split into two superdialects, namely, an urban speech used
across major American and Spanish citites and a diverse form that encompasses
rural areas and small towns. The latter can be further clustered into smaller
varieties with a stronger regional character.
| no_new_dataset | 0.799364 |
1410.0745 | Robinson Piramuthu Robinson Piramuthu | Qiaosong Wang, Vignesh Jagadeesh, Bryan Ressler, Robinson Piramuthu | Im2Fit: Fast 3D Model Fitting and Anthropometrics using Single Consumer
Depth Camera and Synthetic Data | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in consumer depth sensors have created many opportunities for
human body measurement and modeling. Estimation of 3D body shape is
particularly useful for fashion e-commerce applications such as virtual try-on
or fit personalization. In this paper, we propose a method for capturing
accurate human body shape and anthropometrics from a single consumer grade
depth sensor. We first generate a large dataset of synthetic 3D human body
models using real-world body size distributions. Next, we estimate key body
measurements from a single monocular depth image. We combine body measurement
estimates with local geometry features around key joint positions to form a
robust multi-dimensional feature vector. This allows us to conduct a fast
nearest-neighbor search to every sample in the dataset and return the closest
one. Compared to existing methods, our approach is able to predict accurate
full body parameters from a partial view using measurement parameters learned
from the synthetic dataset. Furthermore, our system is capable of generating 3D
human mesh models in real-time, which is significantly faster than methods
which attempt to model shape and pose deformations. To validate the efficiency
and applicability of our system, we collected a dataset that contains frontal
and back scans of 83 clothed people with ground truth height and weight.
Experiments on real-world dataset show that the proposed method can achieve
real-time performance with competing results achieving an average error of 1.9
cm in estimated measurements.
| [
{
"version": "v1",
"created": "Fri, 3 Oct 2014 02:33:08 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Nov 2014 20:30:32 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Wang",
"Qiaosong",
""
],
[
"Jagadeesh",
"Vignesh",
""
],
[
"Ressler",
"Bryan",
""
],
[
"Piramuthu",
"Robinson",
""
]
] | TITLE: Im2Fit: Fast 3D Model Fitting and Anthropometrics using Single Consumer
Depth Camera and Synthetic Data
ABSTRACT: Recent advances in consumer depth sensors have created many opportunities for
human body measurement and modeling. Estimation of 3D body shape is
particularly useful for fashion e-commerce applications such as virtual try-on
or fit personalization. In this paper, we propose a method for capturing
accurate human body shape and anthropometrics from a single consumer grade
depth sensor. We first generate a large dataset of synthetic 3D human body
models using real-world body size distributions. Next, we estimate key body
measurements from a single monocular depth image. We combine body measurement
estimates with local geometry features around key joint positions to form a
robust multi-dimensional feature vector. This allows us to conduct a fast
nearest-neighbor search to every sample in the dataset and return the closest
one. Compared to existing methods, our approach is able to predict accurate
full body parameters from a partial view using measurement parameters learned
from the synthetic dataset. Furthermore, our system is capable of generating 3D
human mesh models in real-time, which is significantly faster than methods
which attempt to model shape and pose deformations. To validate the efficiency
and applicability of our system, we collected a dataset that contains frontal
and back scans of 83 clothed people with ground truth height and weight.
Experiments on real-world dataset show that the proposed method can achieve
real-time performance with competing results achieving an average error of 1.9
cm in estimated measurements.
| new_dataset | 0.958538 |
1411.5140 | Qian Wang | Qian Wang, Jiaxing Zhang, Sen Song, Zheng Zhang | Attentional Neural Network: Feature Selection Using Cognitive Feedback | Poster in Neural Information Processing Systems (NIPS) 2014 | null | null | null | cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Attentional Neural Network is a new framework that integrates top-down
cognitive bias and bottom-up feature extraction in one coherent architecture.
The top-down influence is especially effective when dealing with high noise or
difficult segmentation problems. Our system is modular and extensible. It is
also easy to train and cheap to run, and yet can accommodate complex behaviors.
We obtain classification accuracy better than or competitive with state of art
results on the MNIST variation dataset, and successfully disentangle overlaid
digits with high success rates. We view such a general purpose framework as an
essential foundation for a larger system emulating the cognitive abilities of
the whole brain.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 08:33:28 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Wang",
"Qian",
""
],
[
"Zhang",
"Jiaxing",
""
],
[
"Song",
"Sen",
""
],
[
"Zhang",
"Zheng",
""
]
] | TITLE: Attentional Neural Network: Feature Selection Using Cognitive Feedback
ABSTRACT: Attentional Neural Network is a new framework that integrates top-down
cognitive bias and bottom-up feature extraction in one coherent architecture.
The top-down influence is especially effective when dealing with high noise or
difficult segmentation problems. Our system is modular and extensible. It is
also easy to train and cheap to run, and yet can accommodate complex behaviors.
We obtain classification accuracy better than or competitive with state of art
results on the MNIST variation dataset, and successfully disentangle overlaid
digits with high success rates. We view such a general purpose framework as an
essential foundation for a larger system emulating the cognitive abilities of
the whole brain.
| no_new_dataset | 0.9462 |
1411.5204 | Diego Saez-Trumper | Alessandro Venerandi, Giovanni Quattrone, Licia Capra, Daniele
Quercia, Diego Saez-Trumper | Measuring Urban Deprivation from User Generated Content | CSCW'15, March 14 - 18 2015, Vancouver, BC, Canada | null | 10.1145/2675133.2675233 | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Measuring socioeconomic deprivation of cities in an accurate and timely
fashion has become a priority for governments around the world, as the massive
urbanization process we are witnessing is causing high levels of inequalities
which require intervention. Traditionally, deprivation indexes have been
derived from census data, which is however very expensive to obtain, and thus
acquired only every few years. Alternative computational methods have been
proposed in recent years to automatically extract proxies of deprivation at a
fine spatio-temporal level of granularity; however, they usually require access
to datasets (e.g., call details records) that are not publicly available to
governments and agencies.
To remedy this, we propose a new method to automatically mine deprivation at
a fine level of spatio-temporal granularity that only requires access to freely
available user-generated content. More precisely, the method needs access to
datasets describing what urban elements are present in the physical
environment; examples of such datasets are Foursquare and OpenStreetMap. Using
these datasets, we quantitatively describe neighborhoods by means of a metric,
called {\em Offering Advantage}, that reflects which urban elements are
distinctive features of each neighborhood. We then use that metric to {\em (i)}
build accurate classifiers of urban deprivation and {\em (ii)} interpret the
outcomes through thematic analysis. We apply the method to three UK urban areas
of different scale and elaborate on the results in terms of precision and
recall.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 12:44:12 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Venerandi",
"Alessandro",
""
],
[
"Quattrone",
"Giovanni",
""
],
[
"Capra",
"Licia",
""
],
[
"Quercia",
"Daniele",
""
],
[
"Saez-Trumper",
"Diego",
""
]
] | TITLE: Measuring Urban Deprivation from User Generated Content
ABSTRACT: Measuring socioeconomic deprivation of cities in an accurate and timely
fashion has become a priority for governments around the world, as the massive
urbanization process we are witnessing is causing high levels of inequalities
which require intervention. Traditionally, deprivation indexes have been
derived from census data, which is however very expensive to obtain, and thus
acquired only every few years. Alternative computational methods have been
proposed in recent years to automatically extract proxies of deprivation at a
fine spatio-temporal level of granularity; however, they usually require access
to datasets (e.g., call details records) that are not publicly available to
governments and agencies.
To remedy this, we propose a new method to automatically mine deprivation at
a fine level of spatio-temporal granularity that only requires access to freely
available user-generated content. More precisely, the method needs access to
datasets describing what urban elements are present in the physical
environment; examples of such datasets are Foursquare and OpenStreetMap. Using
these datasets, we quantitatively describe neighborhoods by means of a metric,
called {\em Offering Advantage}, that reflects which urban elements are
distinctive features of each neighborhood. We then use that metric to {\em (i)}
build accurate classifiers of urban deprivation and {\em (ii)} interpret the
outcomes through thematic analysis. We apply the method to three UK urban areas
of different scale and elaborate on the results in terms of precision and
recall.
| no_new_dataset | 0.950134 |
1411.5260 | Patrick Kimes | Patrick K. Kimes, D. Neil Hayes, J. S. Marron and Yufeng Liu | Large-Margin Classification with Multiple Decision Rules | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Binary classification is a common statistical learning problem in which a
model is estimated on a set of covariates for some outcome indicating the
membership of one of two classes. In the literature, there exists a distinction
between hard and soft classification. In soft classification, the conditional
class probability is modeled as a function of the covariates. In contrast, hard
classification methods only target the optimal prediction boundary. While hard
and soft classification methods have been studied extensively, not much work
has been done to compare the actual tasks of hard and soft classification. In
this paper we propose a spectrum of statistical learning problems which span
the hard and soft classification tasks based on fitting multiple decision rules
to the data. By doing so, we reveal a novel collection of learning tasks of
increasing complexity. We study the problems using the framework of
large-margin classifiers and a class of piecewise linear convex surrogates, for
which we derive statistical properties and a corresponding sub-gradient descent
algorithm. We conclude by applying our approach to simulation settings and a
magnetic resonance imaging (MRI) dataset from the Alzheimer's Disease
Neuroimaging Initiative (ADNI) study.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 15:45:54 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Kimes",
"Patrick K.",
""
],
[
"Hayes",
"D. Neil",
""
],
[
"Marron",
"J. S.",
""
],
[
"Liu",
"Yufeng",
""
]
] | TITLE: Large-Margin Classification with Multiple Decision Rules
ABSTRACT: Binary classification is a common statistical learning problem in which a
model is estimated on a set of covariates for some outcome indicating the
membership of one of two classes. In the literature, there exists a distinction
between hard and soft classification. In soft classification, the conditional
class probability is modeled as a function of the covariates. In contrast, hard
classification methods only target the optimal prediction boundary. While hard
and soft classification methods have been studied extensively, not much work
has been done to compare the actual tasks of hard and soft classification. In
this paper we propose a spectrum of statistical learning problems which span
the hard and soft classification tasks based on fitting multiple decision rules
to the data. By doing so, we reveal a novel collection of learning tasks of
increasing complexity. We study the problems using the framework of
large-margin classifiers and a class of piecewise linear convex surrogates, for
which we derive statistical properties and a corresponding sub-gradient descent
algorithm. We conclude by applying our approach to simulation settings and a
magnetic resonance imaging (MRI) dataset from the Alzheimer's Disease
Neuroimaging Initiative (ADNI) study.
| no_new_dataset | 0.942401 |
1411.5283 | Zaid Alyasseri | Zaid Abdi Alkareem Alyasseri, Kadhim Al-Attar, Mazin Nasser | Parallelize Bubble and Merge Sort Algorithms Using Message Passing
Interface (MPI) | 5 pages, 5 figures. arXiv admin note: substantial text overlap with
arXiv:1407.6603 | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sorting has been a profound area for the algorithmic researchers and many
resources are invested to suggest more works for sorting algorithms. For this
purpose, many existing sorting algorithms were observed in terms of the
efficiency of the algorithmic complexity. In this paper we implemented the
bubble and merge sort algorithms using Message Passing Interface (MPI)
approach. The proposed work tested on two standard datasets (text file) with
different size. The main idea of the proposed algorithm is distributing the
elements of the input datasets into many additional temporary sub-arrays
according to a number of characters in each word. The sizes of each of these
sub-arrays are decided depending on a number of elements with the same number
of characters in the input array. We implemented MPI using Intel core i7-3610QM
,(8 CPUs),using two approaches (vectors of string and array 3D) . Finally, we
get the data structure effects on the performance of the algorithm for that we
choice the second approach.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 16:35:16 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Alyasseri",
"Zaid Abdi Alkareem",
""
],
[
"Al-Attar",
"Kadhim",
""
],
[
"Nasser",
"Mazin",
""
]
] | TITLE: Parallelize Bubble and Merge Sort Algorithms Using Message Passing
Interface (MPI)
ABSTRACT: Sorting has been a profound area for the algorithmic researchers and many
resources are invested to suggest more works for sorting algorithms. For this
purpose, many existing sorting algorithms were observed in terms of the
efficiency of the algorithmic complexity. In this paper we implemented the
bubble and merge sort algorithms using Message Passing Interface (MPI)
approach. The proposed work tested on two standard datasets (text file) with
different size. The main idea of the proposed algorithm is distributing the
elements of the input datasets into many additional temporary sub-arrays
according to a number of characters in each word. The sizes of each of these
sub-arrays are decided depending on a number of elements with the same number
of characters in the input array. We implemented MPI using Intel core i7-3610QM
,(8 CPUs),using two approaches (vectors of string and array 3D) . Finally, we
get the data structure effects on the performance of the algorithm for that we
choice the second approach.
| no_new_dataset | 0.949106 |
1411.5307 | Robinson Piramuthu Robinson Piramuthu | Kevin Shih, Wei Di, Vignesh Jagadeesh, Robinson Piramuthu | Efficient Media Retrieval from Non-Cooperative Queries | 8 pages, 9 figures, 1 table | null | null | null | cs.IR cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text is ubiquitous in the artificial world and easily attainable when it
comes to book title and author names. Using the images from the book cover set
from the Stanford Mobile Visual Search dataset and additional book covers and
metadata from openlibrary.org, we construct a large scale book cover retrieval
dataset, complete with 100K distractor covers and title and author strings for
each. Because our query images are poorly conditioned for clean text
extraction, we propose a method for extracting a matching noisy and erroneous
OCR readings and matching it against clean author and book title strings in a
standard document look-up problem setup. Finally, we demonstrate how to use
this text-matching as a feature in conjunction with popular retrieval features
such as VLAD using a simple learning setup to achieve significant improvements
in retrieval accuracy over that of either VLAD or the text alone.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 18:34:28 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Shih",
"Kevin",
""
],
[
"Di",
"Wei",
""
],
[
"Jagadeesh",
"Vignesh",
""
],
[
"Piramuthu",
"Robinson",
""
]
] | TITLE: Efficient Media Retrieval from Non-Cooperative Queries
ABSTRACT: Text is ubiquitous in the artificial world and easily attainable when it
comes to book title and author names. Using the images from the book cover set
from the Stanford Mobile Visual Search dataset and additional book covers and
metadata from openlibrary.org, we construct a large scale book cover retrieval
dataset, complete with 100K distractor covers and title and author strings for
each. Because our query images are poorly conditioned for clean text
extraction, we propose a method for extracting a matching noisy and erroneous
OCR readings and matching it against clean author and book title strings in a
standard document look-up problem setup. Finally, we demonstrate how to use
this text-matching as a feature in conjunction with popular retrieval features
such as VLAD using a simple learning setup to achieve significant improvements
in retrieval accuracy over that of either VLAD or the text alone.
| new_dataset | 0.959078 |
1411.5309 | David Eigen | Li Wan and David Eigen and Rob Fergus | End-to-End Integration of a Convolutional Network, Deformable Parts
Model and Non-Maximum Suppression | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deformable Parts Models and Convolutional Networks each have achieved notable
performance in object detection. Yet these two approaches find their strengths
in complementary areas: DPMs are well-versed in object composition, modeling
fine-grained spatial relationships between parts; likewise, ConvNets are adept
at producing powerful image features, having been discriminatively trained
directly on the pixels. In this paper, we propose a new model that combines
these two approaches, obtaining the advantages of each. We train this model
using a new structured loss function that considers all bounding boxes within
an image, rather than isolated object instances. This enables the non-maximal
suppression (NMS) operation, previously treated as a separate post-processing
stage, to be integrated into the model. This allows for discriminative training
of our combined Convnet + DPM + NMS model in end-to-end fashion. We evaluate
our system on PASCAL VOC 2007 and 2011 datasets, achieving competitive results
on both benchmarks.
| [
{
"version": "v1",
"created": "Wed, 19 Nov 2014 18:36:09 GMT"
}
] | 2014-11-20T00:00:00 | [
[
"Wan",
"Li",
""
],
[
"Eigen",
"David",
""
],
[
"Fergus",
"Rob",
""
]
] | TITLE: End-to-End Integration of a Convolutional Network, Deformable Parts
Model and Non-Maximum Suppression
ABSTRACT: Deformable Parts Models and Convolutional Networks each have achieved notable
performance in object detection. Yet these two approaches find their strengths
in complementary areas: DPMs are well-versed in object composition, modeling
fine-grained spatial relationships between parts; likewise, ConvNets are adept
at producing powerful image features, having been discriminatively trained
directly on the pixels. In this paper, we propose a new model that combines
these two approaches, obtaining the advantages of each. We train this model
using a new structured loss function that considers all bounding boxes within
an image, rather than isolated object instances. This enables the non-maximal
suppression (NMS) operation, previously treated as a separate post-processing
stage, to be integrated into the model. This allows for discriminative training
of our combined Convnet + DPM + NMS model in end-to-end fashion. We evaluate
our system on PASCAL VOC 2007 and 2011 datasets, achieving competitive results
on both benchmarks.
| no_new_dataset | 0.949059 |
1402.5138 | Dieter Pfoser | Mahmuda Ahmed and Sophia Karagiorgou and Dieter Pfoser and Carola Wenk | A Comparison and Evaluation of Map Construction Algorithms | null | null | 10.1007/s10707-014-0222-6 | null | cs.CG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Map construction methods automatically produce and/or update road network
datasets using vehicle tracking data. Enabled by the ubiquitous generation of
georeferenced tracking data, there has been a recent surge in map construction
algorithms coming from different computer science domains. A cross-comparison
of the various algorithms is still very rare, since (i) algorithms and
constructed maps are generally not publicly available and (ii) there is no
standard approach to assess the result quality, given the lack of benchmark
data and quantitative evaluation methods. This work represents a first
comprehensive attempt to benchmark map construction algorithms. We provide an
evaluation and comparison of seven algorithms using four datasets and four
different evaluation measures. In addition to this comprehensive comparison, we
make our datasets, source code of map construction algorithms and evaluation
measures publicly available on mapconstruction.org. This site has been
established as a repository for map con- struction data and algorithms and we
invite other researchers to contribute by uploading code and benchmark data
supporting their contributions to map construction algorithms.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2014 21:50:39 GMT"
},
{
"version": "v2",
"created": "Thu, 12 Jun 2014 14:51:46 GMT"
}
] | 2014-11-19T00:00:00 | [
[
"Ahmed",
"Mahmuda",
""
],
[
"Karagiorgou",
"Sophia",
""
],
[
"Pfoser",
"Dieter",
""
],
[
"Wenk",
"Carola",
""
]
] | TITLE: A Comparison and Evaluation of Map Construction Algorithms
ABSTRACT: Map construction methods automatically produce and/or update road network
datasets using vehicle tracking data. Enabled by the ubiquitous generation of
georeferenced tracking data, there has been a recent surge in map construction
algorithms coming from different computer science domains. A cross-comparison
of the various algorithms is still very rare, since (i) algorithms and
constructed maps are generally not publicly available and (ii) there is no
standard approach to assess the result quality, given the lack of benchmark
data and quantitative evaluation methods. This work represents a first
comprehensive attempt to benchmark map construction algorithms. We provide an
evaluation and comparison of seven algorithms using four datasets and four
different evaluation measures. In addition to this comprehensive comparison, we
make our datasets, source code of map construction algorithms and evaluation
measures publicly available on mapconstruction.org. This site has been
established as a repository for map con- struction data and algorithms and we
invite other researchers to contribute by uploading code and benchmark data
supporting their contributions to map construction algorithms.
| no_new_dataset | 0.784195 |
1409.6911 | Zhiqiang Shen | Zhiqiang Shen and Xiangyang Xue | Do More Dropouts in Pool5 Feature Maps for Better Object Detection | 9 pages, 7 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep Convolutional Neural Networks (CNNs) have gained great success in image
classification and object detection. In these fields, the outputs of all layers
of CNNs are usually considered as a high dimensional feature vector extracted
from an input image and the correspondence between finer level feature vectors
and concepts that the input image contains is all-important. However, fewer
studies focus on this deserving issue. On considering the correspondence, we
propose a novel approach which generates an edited version for each original
CNN feature vector by applying the maximum entropy principle to abandon
particular vectors. These selected vectors correspond to the unfriendly
concepts in each image category. The classifier trained from merged feature
sets can significantly improve model generalization of individual categories
when training data is limited. The experimental results for
classification-based object detection on canonical datasets including VOC 2007
(60.1%), 2010 (56.4%) and 2012 (56.3%) show obvious improvement in mean average
precision (mAP) with simple linear support vector machines.
| [
{
"version": "v1",
"created": "Wed, 24 Sep 2014 11:50:48 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Nov 2014 12:27:23 GMT"
},
{
"version": "v3",
"created": "Tue, 18 Nov 2014 17:34:22 GMT"
}
] | 2014-11-19T00:00:00 | [
[
"Shen",
"Zhiqiang",
""
],
[
"Xue",
"Xiangyang",
""
]
] | TITLE: Do More Dropouts in Pool5 Feature Maps for Better Object Detection
ABSTRACT: Deep Convolutional Neural Networks (CNNs) have gained great success in image
classification and object detection. In these fields, the outputs of all layers
of CNNs are usually considered as a high dimensional feature vector extracted
from an input image and the correspondence between finer level feature vectors
and concepts that the input image contains is all-important. However, fewer
studies focus on this deserving issue. On considering the correspondence, we
propose a novel approach which generates an edited version for each original
CNN feature vector by applying the maximum entropy principle to abandon
particular vectors. These selected vectors correspond to the unfriendly
concepts in each image category. The classifier trained from merged feature
sets can significantly improve model generalization of individual categories
when training data is limited. The experimental results for
classification-based object detection on canonical datasets including VOC 2007
(60.1%), 2010 (56.4%) and 2012 (56.3%) show obvious improvement in mean average
precision (mAP) with simple linear support vector machines.
| no_new_dataset | 0.950549 |
1411.4670 | Mohamed Hussein | Mohamed E. Hussein and Marwan Torki and Ahmed Elsallamy and Mahmoud
Fayyaz | AlexU-Word: A New Dataset for Isolated-Word Closed-Vocabulary Offline
Arabic Handwriting Recognition | 6 pages, 8 figure, and 6 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce the first phase of a new dataset for offline
Arabic handwriting recognition. The aim is to collect a very large dataset of
isolated Arabic words that covers all letters of the alphabet in all possible
shapes using a small number of simple words. The end goal is to collect a very
large dataset of segmented letter images, which can be used to build and
evaluate Arabic handwriting recognition systems that are based on segmented
letter recognition. The current version of the dataset contains $25114$ samples
of $109$ unique Arabic words that cover all possible shapes of all alphabet
letters. The samples were collected from $907$ writers. In its current form,
the dataset can be used for the problem of closed-vocabulary word recognition.
We evaluated a number of window-based descriptors and classifiers on this task
and obtained an accuracy of $92.16\%$ using a SIFT-based descriptor and ANN.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 21:23:26 GMT"
}
] | 2014-11-19T00:00:00 | [
[
"Hussein",
"Mohamed E.",
""
],
[
"Torki",
"Marwan",
""
],
[
"Elsallamy",
"Ahmed",
""
],
[
"Fayyaz",
"Mahmoud",
""
]
] | TITLE: AlexU-Word: A New Dataset for Isolated-Word Closed-Vocabulary Offline
Arabic Handwriting Recognition
ABSTRACT: In this paper, we introduce the first phase of a new dataset for offline
Arabic handwriting recognition. The aim is to collect a very large dataset of
isolated Arabic words that covers all letters of the alphabet in all possible
shapes using a small number of simple words. The end goal is to collect a very
large dataset of segmented letter images, which can be used to build and
evaluate Arabic handwriting recognition systems that are based on segmented
letter recognition. The current version of the dataset contains $25114$ samples
of $109$ unique Arabic words that cover all possible shapes of all alphabet
letters. The samples were collected from $907$ writers. In its current form,
the dataset can be used for the problem of closed-vocabulary word recognition.
We evaluated a number of window-based descriptors and classifiers on this task
and obtained an accuracy of $92.16\%$ using a SIFT-based descriptor and ANN.
| new_dataset | 0.956227 |
1411.4958 | Xiaolong Wang | Xiaolong Wang, David F. Fouhey, Abhinav Gupta | Designing Deep Networks for Surface Normal Estimation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the past few years, convolutional neural nets (CNN) have shown incredible
promise for learning visual representations. In this paper, we use CNNs for the
task of predicting surface normals from a single image. But what is the right
architecture we should use? We propose to build upon the decades of hard work
in 3D scene understanding, to design new CNN architecture for the task of
surface normal estimation. We show by incorporating several constraints
(man-made, manhattan world) and meaningful intermediate representations (room
layout, edge labels) in the architecture leads to state of the art performance
on surface normal estimation. We also show that our network is quite robust and
show state of the art results on other datasets as well without any
fine-tuning.
| [
{
"version": "v1",
"created": "Tue, 18 Nov 2014 18:39:48 GMT"
}
] | 2014-11-19T00:00:00 | [
[
"Wang",
"Xiaolong",
""
],
[
"Fouhey",
"David F.",
""
],
[
"Gupta",
"Abhinav",
""
]
] | TITLE: Designing Deep Networks for Surface Normal Estimation
ABSTRACT: In the past few years, convolutional neural nets (CNN) have shown incredible
promise for learning visual representations. In this paper, we use CNNs for the
task of predicting surface normals from a single image. But what is the right
architecture we should use? We propose to build upon the decades of hard work
in 3D scene understanding, to design new CNN architecture for the task of
surface normal estimation. We show by incorporating several constraints
(man-made, manhattan world) and meaningful intermediate representations (room
layout, edge labels) in the architecture leads to state of the art performance
on surface normal estimation. We also show that our network is quite robust and
show state of the art results on other datasets as well without any
fine-tuning.
| no_new_dataset | 0.955026 |
1411.4960 | Ana Me\v{s}trovi\'c | Hana Rizvi\'c, Sanda Martin\v{c}i\'c-Ip\v{s}i\'c, Ana Me\v{s}trovi\'c | Network Motifs Analysis of Croatian Literature | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we analyse network motifs in the co-occurrence directed
networks constructed from five different texts (four books and one portal) in
the Croatian language. After preparing the data and network construction, we
perform the network motif analysis. We analyse the motif frequencies and
Z-scores in the five networks. We present the triad significance profile for
five datasets. Furthermore, we compare our results with the existing results
for the linguistic networks. Firstly, we show that the triad significance
profile for the Croatian language is very similar with the other languages and
all the networks belong to the same family of networks. However, there are
certain differences between the Croatian language and other analysed languages.
We conclude that this is due to the free word-order of the Croatian language.
| [
{
"version": "v1",
"created": "Tue, 18 Nov 2014 18:46:36 GMT"
}
] | 2014-11-19T00:00:00 | [
[
"Rizvić",
"Hana",
""
],
[
"Martinčić-Ipšić",
"Sanda",
""
],
[
"Meštrović",
"Ana",
""
]
] | TITLE: Network Motifs Analysis of Croatian Literature
ABSTRACT: In this paper we analyse network motifs in the co-occurrence directed
networks constructed from five different texts (four books and one portal) in
the Croatian language. After preparing the data and network construction, we
perform the network motif analysis. We analyse the motif frequencies and
Z-scores in the five networks. We present the triad significance profile for
five datasets. Furthermore, we compare our results with the existing results
for the linguistic networks. Firstly, we show that the triad significance
profile for the Croatian language is very similar with the other languages and
all the networks belong to the same family of networks. However, there are
certain differences between the Croatian language and other analysed languages.
We conclude that this is due to the free word-order of the Croatian language.
| no_new_dataset | 0.946745 |
1403.3155 | Fei-Yun Zhu | Feiyun Zhu, Ying Wang, Bin Fan, Gaofeng Meng, Shiming Xiang and
Chunhong Pan | Spectral Unmixing via Data-guided Sparsity | null | null | 10.1109/TIP.2014.2363423 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hyperspectral unmixing, the process of estimating a common set of spectral
bases and their corresponding composite percentages at each pixel, is an
important task for hyperspectral analysis, visualization and understanding.
From an unsupervised learning perspective, this problem is very
challenging---both the spectral bases and their composite percentages are
unknown, making the solution space too large. To reduce the solution space,
many approaches have been proposed by exploiting various priors. In practice,
these priors would easily lead to some unsuitable solution. This is because
they are achieved by applying an identical strength of constraints to all the
factors, which does not hold in practice. To overcome this limitation, we
propose a novel sparsity based method by learning a data-guided map to describe
the individual mixed level of each pixel. Through this data-guided map, the
$\ell_{p}(0<p<1)$ constraint is applied in an adaptive manner. Such
implementation not only meets the practical situation, but also guides the
spectral bases toward the pixels under highly sparse constraint. What's more,
an elegant optimization scheme as well as its convergence proof have been
provided in this paper. Extensive experiments on several datasets also
demonstrate that the data-guided map is feasible, and high quality unmixing
results could be obtained by our method.
| [
{
"version": "v1",
"created": "Thu, 13 Mar 2014 03:29:22 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Jul 2014 13:49:14 GMT"
},
{
"version": "v3",
"created": "Fri, 19 Sep 2014 02:59:51 GMT"
},
{
"version": "v4",
"created": "Mon, 17 Nov 2014 15:18:15 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Zhu",
"Feiyun",
""
],
[
"Wang",
"Ying",
""
],
[
"Fan",
"Bin",
""
],
[
"Meng",
"Gaofeng",
""
],
[
"Xiang",
"Shiming",
""
],
[
"Pan",
"Chunhong",
""
]
] | TITLE: Spectral Unmixing via Data-guided Sparsity
ABSTRACT: Hyperspectral unmixing, the process of estimating a common set of spectral
bases and their corresponding composite percentages at each pixel, is an
important task for hyperspectral analysis, visualization and understanding.
From an unsupervised learning perspective, this problem is very
challenging---both the spectral bases and their composite percentages are
unknown, making the solution space too large. To reduce the solution space,
many approaches have been proposed by exploiting various priors. In practice,
these priors would easily lead to some unsuitable solution. This is because
they are achieved by applying an identical strength of constraints to all the
factors, which does not hold in practice. To overcome this limitation, we
propose a novel sparsity based method by learning a data-guided map to describe
the individual mixed level of each pixel. Through this data-guided map, the
$\ell_{p}(0<p<1)$ constraint is applied in an adaptive manner. Such
implementation not only meets the practical situation, but also guides the
spectral bases toward the pixels under highly sparse constraint. What's more,
an elegant optimization scheme as well as its convergence proof have been
provided in this paper. Extensive experiments on several datasets also
demonstrate that the data-guided map is feasible, and high quality unmixing
results could be obtained by our method.
| no_new_dataset | 0.948298 |
1407.4764 | Ken Chatfield | Ken Chatfield, Karen Simonyan and Andrew Zisserman | Efficient On-the-fly Category Retrieval using ConvNets and GPUs | Published in proceedings of ACCV 2014 | null | null | null | cs.CV cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the gains in precision and speed, that can be obtained by
using Convolutional Networks (ConvNets) for on-the-fly retrieval - where
classifiers are learnt at run time for a textual query from downloaded images,
and used to rank large image or video datasets.
We make three contributions: (i) we present an evaluation of state-of-the-art
image representations for object category retrieval over standard benchmark
datasets containing 1M+ images; (ii) we show that ConvNets can be used to
obtain features which are incredibly performant, and yet much lower dimensional
than previous state-of-the-art image representations, and that their
dimensionality can be reduced further without loss in performance by
compression using product quantization or binarization. Consequently, features
with the state-of-the-art performance on large-scale datasets of millions of
images can fit in the memory of even a commodity GPU card; (iii) we show that
an SVM classifier can be learnt within a ConvNet framework on a GPU in parallel
with downloading the new training images, allowing for a continuous refinement
of the model as more images become available, and simultaneous training and
ranking. The outcome is an on-the-fly system that significantly outperforms its
predecessors in terms of: precision of retrieval, memory requirements, and
speed, facilitating accurate on-the-fly learning and ranking in under a second
on a single GPU.
| [
{
"version": "v1",
"created": "Thu, 17 Jul 2014 18:29:38 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Nov 2014 08:27:23 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Nov 2014 12:10:23 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Chatfield",
"Ken",
""
],
[
"Simonyan",
"Karen",
""
],
[
"Zisserman",
"Andrew",
""
]
] | TITLE: Efficient On-the-fly Category Retrieval using ConvNets and GPUs
ABSTRACT: We investigate the gains in precision and speed, that can be obtained by
using Convolutional Networks (ConvNets) for on-the-fly retrieval - where
classifiers are learnt at run time for a textual query from downloaded images,
and used to rank large image or video datasets.
We make three contributions: (i) we present an evaluation of state-of-the-art
image representations for object category retrieval over standard benchmark
datasets containing 1M+ images; (ii) we show that ConvNets can be used to
obtain features which are incredibly performant, and yet much lower dimensional
than previous state-of-the-art image representations, and that their
dimensionality can be reduced further without loss in performance by
compression using product quantization or binarization. Consequently, features
with the state-of-the-art performance on large-scale datasets of millions of
images can fit in the memory of even a commodity GPU card; (iii) we show that
an SVM classifier can be learnt within a ConvNet framework on a GPU in parallel
with downloading the new training images, allowing for a continuous refinement
of the model as more images become available, and simultaneous training and
ranking. The outcome is an on-the-fly system that significantly outperforms its
predecessors in terms of: precision of retrieval, memory requirements, and
speed, facilitating accurate on-the-fly learning and ranking in under a second
on a single GPU.
| no_new_dataset | 0.948394 |
1409.3660 | Fei-Yun Zhu | Feiyun Zhu, Bin Fan, Xinliang Zhu, Ying Wang, Shiming Xiang and
Chunhong Pan | 10,000+ Times Accelerated Robust Subset Selection (ARSS) | null | null | null | null | cs.LG cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Subset selection from massive data with noised information is increasingly
popular for various applications. This problem is still highly challenging as
current methods are generally slow in speed and sensitive to outliers. To
address the above two issues, we propose an accelerated robust subset selection
(ARSS) method. Specifically in the subset selection area, this is the first
attempt to employ the $\ell_{p}(0<p\leq1)$-norm based measure for the
representation loss, preventing large errors from dominating our objective. As
a result, the robustness against outlier elements is greatly enhanced.
Actually, data size is generally much larger than feature length, i.e. $N\gg
L$. Based on this observation, we propose a speedup solver (via ALM and
equivalent derivations) to highly reduce the computational cost, theoretically
from $O(N^{4})$ to $O(N{}^{2}L)$. Extensive experiments on ten benchmark
datasets verify that our method not only outperforms state of the art methods,
but also runs 10,000+ times faster than the most related method.
| [
{
"version": "v1",
"created": "Fri, 12 Sep 2014 07:18:17 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Sep 2014 02:49:19 GMT"
},
{
"version": "v3",
"created": "Mon, 13 Oct 2014 07:58:57 GMT"
},
{
"version": "v4",
"created": "Mon, 17 Nov 2014 14:39:31 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Zhu",
"Feiyun",
""
],
[
"Fan",
"Bin",
""
],
[
"Zhu",
"Xinliang",
""
],
[
"Wang",
"Ying",
""
],
[
"Xiang",
"Shiming",
""
],
[
"Pan",
"Chunhong",
""
]
] | TITLE: 10,000+ Times Accelerated Robust Subset Selection (ARSS)
ABSTRACT: Subset selection from massive data with noised information is increasingly
popular for various applications. This problem is still highly challenging as
current methods are generally slow in speed and sensitive to outliers. To
address the above two issues, we propose an accelerated robust subset selection
(ARSS) method. Specifically in the subset selection area, this is the first
attempt to employ the $\ell_{p}(0<p\leq1)$-norm based measure for the
representation loss, preventing large errors from dominating our objective. As
a result, the robustness against outlier elements is greatly enhanced.
Actually, data size is generally much larger than feature length, i.e. $N\gg
L$. Based on this observation, we propose a speedup solver (via ALM and
equivalent derivations) to highly reduce the computational cost, theoretically
from $O(N^{4})$ to $O(N{}^{2}L)$. Extensive experiments on ten benchmark
datasets verify that our method not only outperforms state of the art methods,
but also runs 10,000+ times faster than the most related method.
| no_new_dataset | 0.944125 |
1411.3519 | Mohamed Hussein | Marwan Torki, Mohamed E. Hussein, Ahmed Elsallamy, Mahmoud Fayyaz,
Shehab Yaser | Window-Based Descriptors for Arabic Handwritten Alphabet Recognition: A
Comparative Study on a Novel Dataset | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a comparative study for window-based descriptors on the
application of Arabic handwritten alphabet recognition. We show a detailed
experimental evaluation of different descriptors with several classifiers. The
objective of the paper is to evaluate different window-based descriptors on the
problem of Arabic letter recognition. Our experiments clearly show that they
perform very well. Moreover, we introduce a novel spatial pyramid partitioning
scheme that enhances the recognition accuracy for most descriptors. In
addition, we introduce a novel dataset for Arabic handwritten isolated alphabet
letters, which can serve as a benchmark for future research.
| [
{
"version": "v1",
"created": "Thu, 13 Nov 2014 12:22:57 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Nov 2014 17:55:32 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Torki",
"Marwan",
""
],
[
"Hussein",
"Mohamed E.",
""
],
[
"Elsallamy",
"Ahmed",
""
],
[
"Fayyaz",
"Mahmoud",
""
],
[
"Yaser",
"Shehab",
""
]
] | TITLE: Window-Based Descriptors for Arabic Handwritten Alphabet Recognition: A
Comparative Study on a Novel Dataset
ABSTRACT: This paper presents a comparative study for window-based descriptors on the
application of Arabic handwritten alphabet recognition. We show a detailed
experimental evaluation of different descriptors with several classifiers. The
objective of the paper is to evaluate different window-based descriptors on the
problem of Arabic letter recognition. Our experiments clearly show that they
perform very well. Moreover, we introduce a novel spatial pyramid partitioning
scheme that enhances the recognition accuracy for most descriptors. In
addition, we introduce a novel dataset for Arabic handwritten isolated alphabet
letters, which can serve as a benchmark for future research.
| new_dataset | 0.961929 |
1411.4080 | Miriam Redi | Miriam Redi, Neil O Hare, Rossano Schifanella, Michele Trevisiol,
Alejandro Jaimes | 6 Seconds of Sound and Vision: Creativity in Micro-Videos | 8 pages, 1 figures, conference IEEE CVPR 2014 | null | 10.1109/CVPR.2014.544 | null | cs.MM cs.CV cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The notion of creativity, as opposed to related concepts such as beauty or
interestingness, has not been studied from the perspective of automatic
analysis of multimedia content. Meanwhile, short online videos shared on social
media platforms, or micro-videos, have arisen as a new medium for creative
expression. In this paper we study creative micro-videos in an effort to
understand the features that make a video creative, and to address the problem
of automatic detection of creative content. Defining creative videos as those
that are novel and have aesthetic value, we conduct a crowdsourcing experiment
to create a dataset of over 3,800 micro-videos labelled as creative and
non-creative. We propose a set of computational features that we map to the
components of our definition of creativity, and conduct an analysis to
determine which of these features correlate most with creative video. Finally,
we evaluate a supervised approach to automatically detect creative video, with
promising results, showing that it is necessary to model both aesthetic value
and novelty to achieve optimal classification accuracy.
| [
{
"version": "v1",
"created": "Fri, 14 Nov 2014 23:29:18 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Redi",
"Miriam",
""
],
[
"Hare",
"Neil O",
""
],
[
"Schifanella",
"Rossano",
""
],
[
"Trevisiol",
"Michele",
""
],
[
"Jaimes",
"Alejandro",
""
]
] | TITLE: 6 Seconds of Sound and Vision: Creativity in Micro-Videos
ABSTRACT: The notion of creativity, as opposed to related concepts such as beauty or
interestingness, has not been studied from the perspective of automatic
analysis of multimedia content. Meanwhile, short online videos shared on social
media platforms, or micro-videos, have arisen as a new medium for creative
expression. In this paper we study creative micro-videos in an effort to
understand the features that make a video creative, and to address the problem
of automatic detection of creative content. Defining creative videos as those
that are novel and have aesthetic value, we conduct a crowdsourcing experiment
to create a dataset of over 3,800 micro-videos labelled as creative and
non-creative. We propose a set of computational features that we map to the
components of our definition of creativity, and conduct an analysis to
determine which of these features correlate most with creative video. Finally,
we evaluate a supervised approach to automatically detect creative video, with
promising results, showing that it is necessary to model both aesthetic value
and novelty to achieve optimal classification accuracy.
| new_dataset | 0.959988 |
1411.4086 | Hongwei Li | Hongwei Li and Bin Yu | Error Rate Bounds and Iterative Weighted Majority Voting for
Crowdsourcing | Journal Submission | null | null | null | stat.ML cs.HC cs.LG math.PR math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Crowdsourcing has become an effective and popular tool for human-powered
computation to label large datasets. Since the workers can be unreliable, it is
common in crowdsourcing to assign multiple workers to one task, and to
aggregate the labels in order to obtain results of high quality. In this paper,
we provide finite-sample exponential bounds on the error rate (in probability
and in expectation) of general aggregation rules under the Dawid-Skene
crowdsourcing model. The bounds are derived for multi-class labeling, and can
be used to analyze many aggregation methods, including majority voting,
weighted majority voting and the oracle Maximum A Posteriori (MAP) rule. We
show that the oracle MAP rule approximately optimizes our upper bound on the
mean error rate of weighted majority voting in certain setting. We propose an
iterative weighted majority voting (IWMV) method that optimizes the error rate
bound and approximates the oracle MAP rule. Its one step version has a provable
theoretical guarantee on the error rate. The IWMV method is intuitive and
computationally simple. Experimental results on simulated and real data show
that IWMV performs at least on par with the state-of-the-art methods, and it
has a much lower computational cost (around one hundred times faster) than the
state-of-the-art methods.
| [
{
"version": "v1",
"created": "Sat, 15 Nov 2014 00:02:34 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Li",
"Hongwei",
""
],
[
"Yu",
"Bin",
""
]
] | TITLE: Error Rate Bounds and Iterative Weighted Majority Voting for
Crowdsourcing
ABSTRACT: Crowdsourcing has become an effective and popular tool for human-powered
computation to label large datasets. Since the workers can be unreliable, it is
common in crowdsourcing to assign multiple workers to one task, and to
aggregate the labels in order to obtain results of high quality. In this paper,
we provide finite-sample exponential bounds on the error rate (in probability
and in expectation) of general aggregation rules under the Dawid-Skene
crowdsourcing model. The bounds are derived for multi-class labeling, and can
be used to analyze many aggregation methods, including majority voting,
weighted majority voting and the oracle Maximum A Posteriori (MAP) rule. We
show that the oracle MAP rule approximately optimizes our upper bound on the
mean error rate of weighted majority voting in certain setting. We propose an
iterative weighted majority voting (IWMV) method that optimizes the error rate
bound and approximates the oracle MAP rule. Its one step version has a provable
theoretical guarantee on the error rate. The IWMV method is intuitive and
computationally simple. Experimental results on simulated and real data show
that IWMV performs at least on par with the state-of-the-art methods, and it
has a much lower computational cost (around one hundred times faster) than the
state-of-the-art methods.
| no_new_dataset | 0.946349 |
1411.4101 | Rahul Mohan Mr. | Rahul Mohan | Deep Deconvolutional Networks for Scene Parsing | null | null | null | null | stat.ML cs.CV cs.LG | http://creativecommons.org/licenses/by/3.0/ | Scene parsing is an important and challenging prob- lem in computer vision.
It requires labeling each pixel in an image with the category it belongs to.
Tradition- ally, it has been approached with hand-engineered features from
color information in images. Recently convolutional neural networks (CNNs),
which automatically learn hierar- chies of features, have achieved record
performance on the task. These approaches typically include a post-processing
technique, such as superpixels, to produce the final label- ing. In this paper,
we propose a novel network architecture that combines deep deconvolutional
neural networks with CNNs. Our experiments show that deconvolutional neu- ral
networks are capable of learning higher order image structure beyond edge
primitives in comparison to CNNs. The new network architecture is employed for
multi-patch training, introduced as part of this work. Multi-patch train- ing
makes it possible to effectively learn spatial priors from scenes. The proposed
approach yields state-of-the-art per- formance on four scene parsing datasets,
namely Stanford Background, SIFT Flow, CamVid, and KITTI. In addition, our
system has the added advantage of having a training system that can be
completely automated end-to-end with- out requiring any post-processing.
| [
{
"version": "v1",
"created": "Sat, 15 Nov 2014 02:03:14 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Mohan",
"Rahul",
""
]
] | TITLE: Deep Deconvolutional Networks for Scene Parsing
ABSTRACT: Scene parsing is an important and challenging prob- lem in computer vision.
It requires labeling each pixel in an image with the category it belongs to.
Tradition- ally, it has been approached with hand-engineered features from
color information in images. Recently convolutional neural networks (CNNs),
which automatically learn hierar- chies of features, have achieved record
performance on the task. These approaches typically include a post-processing
technique, such as superpixels, to produce the final label- ing. In this paper,
we propose a novel network architecture that combines deep deconvolutional
neural networks with CNNs. Our experiments show that deconvolutional neu- ral
networks are capable of learning higher order image structure beyond edge
primitives in comparison to CNNs. The new network architecture is employed for
multi-patch training, introduced as part of this work. Multi-patch train- ing
makes it possible to effectively learn spatial priors from scenes. The proposed
approach yields state-of-the-art per- formance on four scene parsing datasets,
namely Stanford Background, SIFT Flow, CamVid, and KITTI. In addition, our
system has the added advantage of having a training system that can be
completely automated end-to-end with- out requiring any post-processing.
| no_new_dataset | 0.952486 |
1411.4246 | Md Lisul Islam | Md. Lisul Islam, Swakkhar Shatabda and M. Sohel Rahman | GreMuTRRR: A Novel Genetic Algorithm to Solve Distance Geometry Problem
for Protein Structures | Accepted for publication in the 8th International Conference on
Electrical and Computer Engineering (ICECE 2014) | null | null | null | cs.NE cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nuclear Magnetic Resonance (NMR) Spectroscopy is a widely used technique to
predict the native structure of proteins. However, NMR machines are only able
to report approximate and partial distances between pair of atoms. To build the
protein structure one has to solve the Euclidean distance geometry problem
given the incomplete interval distance data produced by NMR machines. In this
paper, we propose a new genetic algorithm for solving the Euclidean distance
geometry problem for protein structure prediction given sparse NMR data. Our
genetic algorithm uses a greedy mutation operator to intensify the search, a
twin removal technique for diversification in the population and a random
restart method to recover stagnation. On a standard set of benchmark dataset,
our algorithm significantly outperforms standard genetic algorithms.
| [
{
"version": "v1",
"created": "Sun, 16 Nov 2014 11:26:06 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Islam",
"Md. Lisul",
""
],
[
"Shatabda",
"Swakkhar",
""
],
[
"Rahman",
"M. Sohel",
""
]
] | TITLE: GreMuTRRR: A Novel Genetic Algorithm to Solve Distance Geometry Problem
for Protein Structures
ABSTRACT: Nuclear Magnetic Resonance (NMR) Spectroscopy is a widely used technique to
predict the native structure of proteins. However, NMR machines are only able
to report approximate and partial distances between pair of atoms. To build the
protein structure one has to solve the Euclidean distance geometry problem
given the incomplete interval distance data produced by NMR machines. In this
paper, we propose a new genetic algorithm for solving the Euclidean distance
geometry problem for protein structure prediction given sparse NMR data. Our
genetic algorithm uses a greedy mutation operator to intensify the search, a
twin removal technique for diversification in the population and a random
restart method to recover stagnation. On a standard set of benchmark dataset,
our algorithm significantly outperforms standard genetic algorithms.
| no_new_dataset | 0.951953 |
1411.4304 | Rodrigo Benenson | Rodrigo Benenson, Mohamed Omran, Jan Hosang, Bernt Schiele | Ten Years of Pedestrian Detection, What Have We Learned? | To appear in ECCV 2014 CVRSUAD workshop proceedings | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Paper-by-paper results make it easy to miss the forest for the trees.We
analyse the remarkable progress of the last decade by discussing the main ideas
explored in the 40+ detectors currently present in the Caltech pedestrian
detection benchmark. We observe that there exist three families of approaches,
all currently reaching similar detection quality. Based on our analysis, we
study the complementarity of the most promising ideas by combining multiple
published strategies. This new decision forest detector achieves the current
best known performance on the challenging Caltech-USA dataset.
| [
{
"version": "v1",
"created": "Sun, 16 Nov 2014 21:25:53 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Benenson",
"Rodrigo",
""
],
[
"Omran",
"Mohamed",
""
],
[
"Hosang",
"Jan",
""
],
[
"Schiele",
"Bernt",
""
]
] | TITLE: Ten Years of Pedestrian Detection, What Have We Learned?
ABSTRACT: Paper-by-paper results make it easy to miss the forest for the trees.We
analyse the remarkable progress of the last decade by discussing the main ideas
explored in the 40+ detectors currently present in the Caltech pedestrian
detection benchmark. We observe that there exist three families of approaches,
all currently reaching similar detection quality. Based on our analysis, we
study the complementarity of the most promising ideas by combining multiple
published strategies. This new decision forest detector achieves the current
best known performance on the challenging Caltech-USA dataset.
| no_new_dataset | 0.946646 |
1411.4314 | Nikolai Sinitsyn | Benjamin H. Sims, Nikolai Sinitsyn, and Stephan J. Eidenbenz | Hierarchical and Matrix Structures in a Large Organizational Email
Network: Visualization and Modeling Approaches | 15 pages, 9 figures | null | null | null | cs.SI physics.soc-ph stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents findings from a study of the email network of a large
scientific research organization, focusing on methods for visualizing and
modeling organizational hierarchies within large, complex network datasets. In
the first part of the paper, we find that visualization and interpretation of
complex organizational network data is facilitated by integration of network
data with information on formal organizational divisions and levels. By
aggregating and visualizing email traffic between organizational units at
various levels, we derive several insights into how large subdivisions of the
organization interact with each other and with outside organizations. Our
analysis shows that line and program management interactions in this
organization systematically deviate from the idealized pattern of interaction
prescribed by "matrix management." In the second part of the paper, we propose
a power law model for predicting degree distribution of organizational email
traffic based on hierarchical relationships between managers and employees.
This model considers the influence of global email announcements sent from
managers to all employees under their supervision, and the role support staff
play in generating email traffic, acting as agents for managers. We also
analyze patterns in email traffic volume over the course of a work week.
| [
{
"version": "v1",
"created": "Sun, 16 Nov 2014 22:22:54 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Sims",
"Benjamin H.",
""
],
[
"Sinitsyn",
"Nikolai",
""
],
[
"Eidenbenz",
"Stephan J.",
""
]
] | TITLE: Hierarchical and Matrix Structures in a Large Organizational Email
Network: Visualization and Modeling Approaches
ABSTRACT: This paper presents findings from a study of the email network of a large
scientific research organization, focusing on methods for visualizing and
modeling organizational hierarchies within large, complex network datasets. In
the first part of the paper, we find that visualization and interpretation of
complex organizational network data is facilitated by integration of network
data with information on formal organizational divisions and levels. By
aggregating and visualizing email traffic between organizational units at
various levels, we derive several insights into how large subdivisions of the
organization interact with each other and with outside organizations. Our
analysis shows that line and program management interactions in this
organization systematically deviate from the idealized pattern of interaction
prescribed by "matrix management." In the second part of the paper, we propose
a power law model for predicting degree distribution of organizational email
traffic based on hierarchical relationships between managers and employees.
This model considers the influence of global email announcements sent from
managers to all employees under their supervision, and the role support staff
play in generating email traffic, acting as agents for managers. We also
analyze patterns in email traffic volume over the course of a work week.
| no_new_dataset | 0.940188 |
1411.4379 | Md Lisul Islam | Md. Lisul Islam, Novia Nurain, Swakkhar Shatabda and M Sohel Rahman | FGPGA: An Efficient Genetic Approach for Producing Feasible Graph
Partitions | Accepted in the 1st International Conference on Networking Systems
and Security 2015 (NSysS 2015) | null | null | null | cs.NE cs.AI cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph partitioning, a well studied problem of parallel computing has many
applications in diversified fields such as distributed computing, social
network analysis, data mining and many other domains. In this paper, we
introduce FGPGA, an efficient genetic approach for producing feasible graph
partitions. Our method takes into account the heterogeneity and capacity
constraints of the partitions to ensure balanced partitioning. Such approach
has various applications in mobile cloud computing that include feasible
deployment of software applications on the more resourceful infrastructure in
the cloud instead of mobile hand set. Our proposed approach is light weight and
hence suitable for use in cloud architecture. We ensure feasibility of the
partitions generated by not allowing over-sized partitions to be generated
during the initialization and search. Our proposed method tested on standard
benchmark datasets significantly outperforms the state-of-the-art methods in
terms of quality of partitions and feasibility of the solutions.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 06:51:50 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Islam",
"Md. Lisul",
""
],
[
"Nurain",
"Novia",
""
],
[
"Shatabda",
"Swakkhar",
""
],
[
"Rahman",
"M Sohel",
""
]
] | TITLE: FGPGA: An Efficient Genetic Approach for Producing Feasible Graph
Partitions
ABSTRACT: Graph partitioning, a well studied problem of parallel computing has many
applications in diversified fields such as distributed computing, social
network analysis, data mining and many other domains. In this paper, we
introduce FGPGA, an efficient genetic approach for producing feasible graph
partitions. Our method takes into account the heterogeneity and capacity
constraints of the partitions to ensure balanced partitioning. Such approach
has various applications in mobile cloud computing that include feasible
deployment of software applications on the more resourceful infrastructure in
the cloud instead of mobile hand set. Our proposed approach is light weight and
hence suitable for use in cloud architecture. We ensure feasibility of the
partitions generated by not allowing over-sized partitions to be generated
during the initialization and search. Our proposed method tested on standard
benchmark datasets significantly outperforms the state-of-the-art methods in
terms of quality of partitions and feasibility of the solutions.
| no_new_dataset | 0.950503 |
1411.4455 | Miao Fan | Miao Fan, Deli Zhao, Qiang Zhou, Zhiyuan Liu, Thomas Fang Zheng,
Edward Y. Chang | Errata: Distant Supervision for Relation Extraction with Matrix
Completion | null | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The essence of distantly supervised relation extraction is that it is an
incomplete multi-label classification problem with sparse and noisy features.
To tackle the sparsity and noise challenges, we propose solving the
classification problem using matrix completion on factorized matrix of
minimized rank. We formulate relation classification as completing the unknown
labels of testing items (entity pairs) in a sparse matrix that concatenates
training and testing textual features with training labels. Our algorithmic
framework is based on the assumption that the rank of item-by-feature and
item-by-label joint matrix is low. We apply two optimization models to recover
the underlying low-rank matrix leveraging the sparsity of feature-label matrix.
The matrix completion problem is then solved by the fixed point continuation
(FPC) algorithm, which can find the global optimum. Experiments on two widely
used datasets with different dimensions of textual features demonstrate that
our low-rank matrix completion approach significantly outperforms the baseline
and the state-of-the-art methods.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 12:43:30 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Fan",
"Miao",
""
],
[
"Zhao",
"Deli",
""
],
[
"Zhou",
"Qiang",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Zheng",
"Thomas Fang",
""
],
[
"Chang",
"Edward Y.",
""
]
] | TITLE: Errata: Distant Supervision for Relation Extraction with Matrix
Completion
ABSTRACT: The essence of distantly supervised relation extraction is that it is an
incomplete multi-label classification problem with sparse and noisy features.
To tackle the sparsity and noise challenges, we propose solving the
classification problem using matrix completion on factorized matrix of
minimized rank. We formulate relation classification as completing the unknown
labels of testing items (entity pairs) in a sparse matrix that concatenates
training and testing textual features with training labels. Our algorithmic
framework is based on the assumption that the rank of item-by-feature and
item-by-label joint matrix is low. We apply two optimization models to recover
the underlying low-rank matrix leveraging the sparsity of feature-label matrix.
The matrix completion problem is then solved by the fixed point continuation
(FPC) algorithm, which can find the global optimum. Experiments on two widely
used datasets with different dimensions of textual features demonstrate that
our low-rank matrix completion approach significantly outperforms the baseline
and the state-of-the-art methods.
| no_new_dataset | 0.939969 |
1411.4464 | Kai Kang | Kai Kang, Xiaogang Wang | Fully Convolutional Neural Networks for Crowd Segmentation | 9 pages,7 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/3.0/ | In this paper, we propose a fast fully convolutional neural network (FCNN)
for crowd segmentation. By replacing the fully connected layers in CNN with 1
by 1 convolution kernels, FCNN takes whole images as inputs and directly
outputs segmentation maps by one pass of forward propagation. It has the
property of translation invariance like patch-by-patch scanning but with much
lower computation cost. Once FCNN is learned, it can process input images of
any sizes without warping them to a standard size. These attractive properties
make it extendable to other general image segmentation problems. Based on FCNN,
a multi-stage deep learning is proposed to integrate appearance and motion cues
for crowd segmentation. Both appearance filters and motion filers are
pretrained stage-by-stage and then jointly optimized. Different combination
methods are investigated. The effectiveness of our approach and component-wise
analysis are evaluated on two crowd segmentation datasets created by us, which
include image frames from 235 and 11 scenes, respectively. They are currently
the largest crowd segmentation datasets and will be released to the public.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 13:09:09 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Kang",
"Kai",
""
],
[
"Wang",
"Xiaogang",
""
]
] | TITLE: Fully Convolutional Neural Networks for Crowd Segmentation
ABSTRACT: In this paper, we propose a fast fully convolutional neural network (FCNN)
for crowd segmentation. By replacing the fully connected layers in CNN with 1
by 1 convolution kernels, FCNN takes whole images as inputs and directly
outputs segmentation maps by one pass of forward propagation. It has the
property of translation invariance like patch-by-patch scanning but with much
lower computation cost. Once FCNN is learned, it can process input images of
any sizes without warping them to a standard size. These attractive properties
make it extendable to other general image segmentation problems. Based on FCNN,
a multi-stage deep learning is proposed to integrate appearance and motion cues
for crowd segmentation. Both appearance filters and motion filers are
pretrained stage-by-stage and then jointly optimized. Different combination
methods are investigated. The effectiveness of our approach and component-wise
analysis are evaluated on two crowd segmentation datasets created by us, which
include image frames from 235 and 11 scenes, respectively. They are currently
the largest crowd segmentation datasets and will be released to the public.
| no_new_dataset | 0.938463 |
1411.4472 | Andrej Gajduk | Andrej Gajduk and Ljupco Kocarev | Opinion mining of text documents written in Macedonian language | In press, MASA proceedings | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ability to extract public opinion from web portals such as review sites,
social networks and blogs will enable companies and individuals to form a view,
an attitude and make decisions without having to do lengthy and costly
researches and surveys. In this paper machine learning techniques are used for
determining the polarity of forum posts on kajgana which are written in
Macedonian language. The posts are classified as being positive, negative or
neutral. We test different feature metrics and classifiers and provide detailed
evaluation of their participation in improving the overall performance on a
manually generated dataset. By achieving 92% accuracy, we show that the
performance of systems for automated opinion mining is comparable to a human
evaluator, thus making it a viable option for text data analysis. Finally, we
present a few statistics derived from the forum posts using the developed
system.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 13:36:49 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Gajduk",
"Andrej",
""
],
[
"Kocarev",
"Ljupco",
""
]
] | TITLE: Opinion mining of text documents written in Macedonian language
ABSTRACT: The ability to extract public opinion from web portals such as review sites,
social networks and blogs will enable companies and individuals to form a view,
an attitude and make decisions without having to do lengthy and costly
researches and surveys. In this paper machine learning techniques are used for
determining the polarity of forum posts on kajgana which are written in
Macedonian language. The posts are classified as being positive, negative or
neutral. We test different feature metrics and classifiers and provide detailed
evaluation of their participation in improving the overall performance on a
manually generated dataset. By achieving 92% accuracy, we show that the
performance of systems for automated opinion mining is comparable to a human
evaluator, thus making it a viable option for text data analysis. Finally, we
present a few statistics derived from the forum posts using the developed
system.
| new_dataset | 0.954774 |
1411.4510 | Kian Hsiang Low | Kian Hsiang Low, Jiangbo Yu, Jie Chen, Patrick Jaillet | Parallel Gaussian Process Regression for Big Data: Low-Rank
Representation Meets Markov Approximation | 29th AAAI Conference on Artificial Intelligence (AAAI 2015), Extended
version with proofs, 10 pages | null | null | null | stat.ML cs.DC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The expressive power of a Gaussian process (GP) model comes at a cost of poor
scalability in the data size. To improve its scalability, this paper presents a
low-rank-cum-Markov approximation (LMA) of the GP model that is novel in
leveraging the dual computational advantages stemming from complementing a
low-rank approximate representation of the full-rank GP based on a support set
of inputs with a Markov approximation of the resulting residual process; the
latter approximation is guaranteed to be closest in the Kullback-Leibler
distance criterion subject to some constraint and is considerably more refined
than that of existing sparse GP models utilizing low-rank representations due
to its more relaxed conditional independence assumption (especially with larger
data). As a result, our LMA method can trade off between the size of the
support set and the order of the Markov property to (a) incur lower
computational cost than such sparse GP models while achieving predictive
performance comparable to them and (b) accurately represent features/patterns
of any scale. Interestingly, varying the Markov order produces a spectrum of
LMAs with PIC approximation and full-rank GP at the two extremes. An advantage
of our LMA method is that it is amenable to parallelization on multiple
machines/cores, thereby gaining greater scalability. Empirical evaluation on
three real-world datasets in clusters of up to 32 computing nodes shows that
our centralized and parallel LMA methods are significantly more time-efficient
and scalable than state-of-the-art sparse and full-rank GP regression methods
while achieving comparable predictive performances.
| [
{
"version": "v1",
"created": "Mon, 17 Nov 2014 15:31:04 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Low",
"Kian Hsiang",
""
],
[
"Yu",
"Jiangbo",
""
],
[
"Chen",
"Jie",
""
],
[
"Jaillet",
"Patrick",
""
]
] | TITLE: Parallel Gaussian Process Regression for Big Data: Low-Rank
Representation Meets Markov Approximation
ABSTRACT: The expressive power of a Gaussian process (GP) model comes at a cost of poor
scalability in the data size. To improve its scalability, this paper presents a
low-rank-cum-Markov approximation (LMA) of the GP model that is novel in
leveraging the dual computational advantages stemming from complementing a
low-rank approximate representation of the full-rank GP based on a support set
of inputs with a Markov approximation of the resulting residual process; the
latter approximation is guaranteed to be closest in the Kullback-Leibler
distance criterion subject to some constraint and is considerably more refined
than that of existing sparse GP models utilizing low-rank representations due
to its more relaxed conditional independence assumption (especially with larger
data). As a result, our LMA method can trade off between the size of the
support set and the order of the Markov property to (a) incur lower
computational cost than such sparse GP models while achieving predictive
performance comparable to them and (b) accurately represent features/patterns
of any scale. Interestingly, varying the Markov order produces a spectrum of
LMAs with PIC approximation and full-rank GP at the two extremes. An advantage
of our LMA method is that it is amenable to parallelization on multiple
machines/cores, thereby gaining greater scalability. Empirical evaluation on
three real-world datasets in clusters of up to 32 computing nodes shows that
our centralized and parallel LMA methods are significantly more time-efficient
and scalable than state-of-the-art sparse and full-rank GP regression methods
while achieving comparable predictive performances.
| no_new_dataset | 0.948251 |
physics/0606042 | Suen Hou | J. Antos, M. Babik, D. Benjamin, S. Cabrera, A.W. Chan, Y.C. Chen, M.
Coca, B. Cooper, S. Farrington, K. Genser, K. Hatakeyama, S. Hou, T.L. Hsieh,
B. Jayatilaka, S.Y. Jun, A.V. Kotwal, A.C. Kraan, R. Lysak, I.V.
Mandrichenko, P. Murat, A. Robson, P. Savard, M. Siket, B. Stelzer, J. Syu,
P.K. Teng, S.C. Timm, T. Tomura, E. Vataga, and S.A. Wolbers | Data processing model for the CDF experiment | 12 pages, 10 figures, submitted to IEEE-TNS | IEEE Trans.Nucl.Sci.53:2897-2906,2006 | 10.1109/TNS.2006.881908 | FERMILAB-PUB-06-169-CD-E | physics.ins-det physics.data-an | null | The data processing model for the CDF experiment is described. Data
processing reconstructs events from parallel data streams taken with different
combinations of physics event triggers and further splits the events into
datasets of specialized physics datasets. The design of the processing control
system faces strict requirements on bookkeeping records, which trace the status
of data files and event contents during processing and storage. The computing
architecture was updated to meet the mass data flow of the Run II data
collection, recently upgraded to a maximum rate of 40 MByte/sec. The data
processing facility consists of a large cluster of Linux computers with data
movement managed by the CDF data handling system to a multi-petaByte Enstore
tape library. The latest processing cycle has achieved a stable speed of 35
MByte/sec (3 TByte/day). It can be readily scaled by increasing CPU and
data-handling capacity as required.
| [
{
"version": "v1",
"created": "Mon, 5 Jun 2006 16:37:59 GMT"
},
{
"version": "v2",
"created": "Fri, 9 Jun 2006 06:23:26 GMT"
}
] | 2014-11-18T00:00:00 | [
[
"Antos",
"J.",
""
],
[
"Babik",
"M.",
""
],
[
"Benjamin",
"D.",
""
],
[
"Cabrera",
"S.",
""
],
[
"Chan",
"A. W.",
""
],
[
"Chen",
"Y. C.",
""
],
[
"Coca",
"M.",
""
],
[
"Cooper",
"B.",
""
],
[
"Farrington",
"S.",
""
],
[
"Genser",
"K.",
""
],
[
"Hatakeyama",
"K.",
""
],
[
"Hou",
"S.",
""
],
[
"Hsieh",
"T. L.",
""
],
[
"Jayatilaka",
"B.",
""
],
[
"Jun",
"S. Y.",
""
],
[
"Kotwal",
"A. V.",
""
],
[
"Kraan",
"A. C.",
""
],
[
"Lysak",
"R.",
""
],
[
"Mandrichenko",
"I. V.",
""
],
[
"Murat",
"P.",
""
],
[
"Robson",
"A.",
""
],
[
"Savard",
"P.",
""
],
[
"Siket",
"M.",
""
],
[
"Stelzer",
"B.",
""
],
[
"Syu",
"J.",
""
],
[
"Teng",
"P. K.",
""
],
[
"Timm",
"S. C.",
""
],
[
"Tomura",
"T.",
""
],
[
"Vataga",
"E.",
""
],
[
"Wolbers",
"S. A.",
""
]
] | TITLE: Data processing model for the CDF experiment
ABSTRACT: The data processing model for the CDF experiment is described. Data
processing reconstructs events from parallel data streams taken with different
combinations of physics event triggers and further splits the events into
datasets of specialized physics datasets. The design of the processing control
system faces strict requirements on bookkeeping records, which trace the status
of data files and event contents during processing and storage. The computing
architecture was updated to meet the mass data flow of the Run II data
collection, recently upgraded to a maximum rate of 40 MByte/sec. The data
processing facility consists of a large cluster of Linux computers with data
movement managed by the CDF data handling system to a multi-petaByte Enstore
tape library. The latest processing cycle has achieved a stable speed of 35
MByte/sec (3 TByte/day). It can be readily scaled by increasing CPU and
data-handling capacity as required.
| no_new_dataset | 0.936168 |
1411.3159 | Marcel Simon | Marcel Simon, Erik Rodner, Joachim Denzler | Part Detector Discovery in Deep Convolutional Neural Networks | Accepted for publication on Asian Conference on Computer Vision
(ACCV) 2014 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current fine-grained classification approaches often rely on a robust
localization of object parts to extract localized feature representations
suitable for discrimination. However, part localization is a challenging task
due to the large variation of appearance and pose. In this paper, we show how
pre-trained convolutional neural networks can be used for robust and efficient
object part discovery and localization without the necessity to actually train
the network on the current dataset. Our approach called "part detector
discovery" (PDD) is based on analyzing the gradient maps of the network outputs
and finding activation centers spatially related to annotated semantic parts or
bounding boxes.
This allows us not just to obtain excellent performance on the CUB200-2011
dataset, but in contrast to previous approaches also to perform detection and
bird classification jointly without requiring a given bounding box annotation
during testing and ground-truth parts during training. The code is available at
http://www.inf-cv.uni-jena.de/part_discovery and
https://github.com/cvjena/PartDetectorDisovery.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 12:42:54 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Nov 2014 11:57:27 GMT"
}
] | 2014-11-17T00:00:00 | [
[
"Simon",
"Marcel",
""
],
[
"Rodner",
"Erik",
""
],
[
"Denzler",
"Joachim",
""
]
] | TITLE: Part Detector Discovery in Deep Convolutional Neural Networks
ABSTRACT: Current fine-grained classification approaches often rely on a robust
localization of object parts to extract localized feature representations
suitable for discrimination. However, part localization is a challenging task
due to the large variation of appearance and pose. In this paper, we show how
pre-trained convolutional neural networks can be used for robust and efficient
object part discovery and localization without the necessity to actually train
the network on the current dataset. Our approach called "part detector
discovery" (PDD) is based on analyzing the gradient maps of the network outputs
and finding activation centers spatially related to annotated semantic parts or
bounding boxes.
This allows us not just to obtain excellent performance on the CUB200-2011
dataset, but in contrast to previous approaches also to perform detection and
bird classification jointly without requiring a given bounding box annotation
during testing and ground-truth parts during training. The code is available at
http://www.inf-cv.uni-jena.de/part_discovery and
https://github.com/cvjena/PartDetectorDisovery.
| no_new_dataset | 0.95511 |
1411.3749 | Timothy La Fond | Timothy La Fond, Jennifer Neville, Brian Gallagher | Anomaly Detection in Dynamic Networks of Varying Size | null | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dynamic networks, also called network streams, are an important data
representation that applies to many real-world domains. Many sets of network
data such as e-mail networks, social networks, or internet traffic networks are
best represented by a dynamic network due to the temporal component of the
data. One important application in the domain of dynamic network analysis is
anomaly detection. Here the task is to identify points in time where the
network exhibits behavior radically different from a typical time, either due
to some event (like the failure of machines in a computer network) or a shift
in the network properties. This problem is made more difficult by the fluid
nature of what is considered "normal" network behavior. The volume of traffic
on a network, for example, can change over the course of a month or even vary
based on the time of the day without being considered unusual. Anomaly
detection tests using traditional network statistics have difficulty in these
scenarios due to their Density Dependence: as the volume of edges changes the
value of the statistics changes as well making it difficult to determine if the
change in signal is due to the traffic volume or due to some fundamental shift
in the behavior of the network. To more accurately detect anomalies in dynamic
networks, we introduce the concept of Density-Consistent network statistics. On
synthetically generated graphs anomaly detectors using these statistics show a
a 20-400% improvement in the recall when distinguishing graphs drawn from
different distributions. When applied to several real datasets
Density-Consistent statistics recover multiple network events which standard
statistics failed to find.
| [
{
"version": "v1",
"created": "Thu, 13 Nov 2014 21:41:55 GMT"
}
] | 2014-11-17T00:00:00 | [
[
"La Fond",
"Timothy",
""
],
[
"Neville",
"Jennifer",
""
],
[
"Gallagher",
"Brian",
""
]
] | TITLE: Anomaly Detection in Dynamic Networks of Varying Size
ABSTRACT: Dynamic networks, also called network streams, are an important data
representation that applies to many real-world domains. Many sets of network
data such as e-mail networks, social networks, or internet traffic networks are
best represented by a dynamic network due to the temporal component of the
data. One important application in the domain of dynamic network analysis is
anomaly detection. Here the task is to identify points in time where the
network exhibits behavior radically different from a typical time, either due
to some event (like the failure of machines in a computer network) or a shift
in the network properties. This problem is made more difficult by the fluid
nature of what is considered "normal" network behavior. The volume of traffic
on a network, for example, can change over the course of a month or even vary
based on the time of the day without being considered unusual. Anomaly
detection tests using traditional network statistics have difficulty in these
scenarios due to their Density Dependence: as the volume of edges changes the
value of the statistics changes as well making it difficult to determine if the
change in signal is due to the traffic volume or due to some fundamental shift
in the behavior of the network. To more accurately detect anomalies in dynamic
networks, we introduce the concept of Density-Consistent network statistics. On
synthetically generated graphs anomaly detectors using these statistics show a
a 20-400% improvement in the recall when distinguishing graphs drawn from
different distributions. When applied to several real datasets
Density-Consistent statistics recover multiple network events which standard
statistics failed to find.
| no_new_dataset | 0.949342 |
1411.3787 | Ping Li | Anshumali Shrivastava, Ping Li | Asymmetric Minwise Hashing | null | null | null | null | stat.ML cs.DB cs.DS cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Minwise hashing (Minhash) is a widely popular indexing scheme in practice.
Minhash is designed for estimating set resemblance and is known to be
suboptimal in many applications where the desired measure is set overlap (i.e.,
inner product between binary vectors) or set containment. Minhash has inherent
bias towards smaller sets, which adversely affects its performance in
applications where such a penalization is not desirable. In this paper, we
propose asymmetric minwise hashing (MH-ALSH), to provide a solution to this
problem. The new scheme utilizes asymmetric transformations to cancel the bias
of traditional minhash towards smaller sets, making the final "collision
probability" monotonic in the inner product. Our theoretical comparisons show
that for the task of retrieving with binary inner products asymmetric minhash
is provably better than traditional minhash and other recently proposed hashing
algorithms for general inner products. Thus, we obtain an algorithmic
improvement over existing approaches in the literature. Experimental
evaluations on four publicly available high-dimensional datasets validate our
claims and the proposed scheme outperforms, often significantly, other hashing
algorithms on the task of near neighbor retrieval with set containment. Our
proposal is simple and easy to implement in practice.
| [
{
"version": "v1",
"created": "Fri, 14 Nov 2014 04:18:33 GMT"
}
] | 2014-11-17T00:00:00 | [
[
"Shrivastava",
"Anshumali",
""
],
[
"Li",
"Ping",
""
]
] | TITLE: Asymmetric Minwise Hashing
ABSTRACT: Minwise hashing (Minhash) is a widely popular indexing scheme in practice.
Minhash is designed for estimating set resemblance and is known to be
suboptimal in many applications where the desired measure is set overlap (i.e.,
inner product between binary vectors) or set containment. Minhash has inherent
bias towards smaller sets, which adversely affects its performance in
applications where such a penalization is not desirable. In this paper, we
propose asymmetric minwise hashing (MH-ALSH), to provide a solution to this
problem. The new scheme utilizes asymmetric transformations to cancel the bias
of traditional minhash towards smaller sets, making the final "collision
probability" monotonic in the inner product. Our theoretical comparisons show
that for the task of retrieving with binary inner products asymmetric minhash
is provably better than traditional minhash and other recently proposed hashing
algorithms for general inner products. Thus, we obtain an algorithmic
improvement over existing approaches in the literature. Experimental
evaluations on four publicly available high-dimensional datasets validate our
claims and the proposed scheme outperforms, often significantly, other hashing
algorithms on the task of near neighbor retrieval with set containment. Our
proposal is simple and easy to implement in practice.
| no_new_dataset | 0.949248 |
1411.4006 | Zhongwen Xu | Zhongwen Xu, Yi Yang and Alexander G. Hauptmann | A Discriminative CNN Video Representation for Event Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose a discriminative video representation for event
detection over a large scale video dataset when only limited hardware resources
are available. The focus of this paper is to effectively leverage deep
Convolutional Neural Networks (CNNs) to advance event detection, where only
frame level static descriptors can be extracted by the existing CNN toolkit.
This paper makes two contributions to the inference of CNN video
representation. First, while average pooling and max pooling have long been the
standard approaches to aggregating frame level static features, we show that
performance can be significantly improved by taking advantage of an appropriate
encoding method. Second, we propose using a set of latent concept descriptors
as the frame descriptor, which enriches visual information while keeping it
computationally affordable. The integration of the two contributions results in
a new state-of-the-art performance in event detection over the largest video
datasets. Compared to improved Dense Trajectories, which has been recognized as
the best video representation for event detection, our new representation
improves the Mean Average Precision (mAP) from 27.6% to 36.8% for the TRECVID
MEDTest 14 dataset and from 34.0% to 44.6% for the TRECVID MEDTest 13 dataset.
This work is the core part of the winning solution of our CMU-Informedia team
in TRECVID MED 2014 competition.
| [
{
"version": "v1",
"created": "Fri, 14 Nov 2014 18:37:31 GMT"
}
] | 2014-11-17T00:00:00 | [
[
"Xu",
"Zhongwen",
""
],
[
"Yang",
"Yi",
""
],
[
"Hauptmann",
"Alexander G.",
""
]
] | TITLE: A Discriminative CNN Video Representation for Event Detection
ABSTRACT: In this paper, we propose a discriminative video representation for event
detection over a large scale video dataset when only limited hardware resources
are available. The focus of this paper is to effectively leverage deep
Convolutional Neural Networks (CNNs) to advance event detection, where only
frame level static descriptors can be extracted by the existing CNN toolkit.
This paper makes two contributions to the inference of CNN video
representation. First, while average pooling and max pooling have long been the
standard approaches to aggregating frame level static features, we show that
performance can be significantly improved by taking advantage of an appropriate
encoding method. Second, we propose using a set of latent concept descriptors
as the frame descriptor, which enriches visual information while keeping it
computationally affordable. The integration of the two contributions results in
a new state-of-the-art performance in event detection over the largest video
datasets. Compared to improved Dense Trajectories, which has been recognized as
the best video representation for event detection, our new representation
improves the Mean Average Precision (mAP) from 27.6% to 36.8% for the TRECVID
MEDTest 14 dataset and from 34.0% to 44.6% for the TRECVID MEDTest 13 dataset.
This work is the core part of the winning solution of our CMU-Informedia team
in TRECVID MED 2014 competition.
| no_new_dataset | 0.948537 |
cs/9508101 | null | Q. Zhao, T. Nishida | Using Qualitative Hypotheses to Identify Inaccurate Data | See http://www.jair.org/ for any accompanying files | Journal of Artificial Intelligence Research, Vol 3, (1995),
119-145 | null | null | cs.AI | null | Identifying inaccurate data has long been regarded as a significant and
difficult problem in AI. In this paper, we present a new method for identifying
inaccurate data on the basis of qualitative correlations among related data.
First, we introduce the definitions of related data and qualitative
correlations among related data. Then we put forward a new concept called
support coefficient function (SCF). SCF can be used to extract, represent, and
calculate qualitative correlations among related data within a dataset. We
propose an approach to determining dynamic shift intervals of inaccurate data,
and an approach to calculating possibility of identifying inaccurate data,
respectively. Both of the approaches are based on SCF. Finally we present an
algorithm for identifying inaccurate data by using qualitative correlations
among related data as confirmatory or disconfirmatory evidence. We have
developed a practical system for interpreting infrared spectra by applying the
method, and have fully tested the system against several hundred real spectra.
The experimental results show that the method is significantly better than the
conventional methods used in many similar systems.
| [
{
"version": "v1",
"created": "Tue, 1 Aug 1995 00:00:00 GMT"
}
] | 2014-11-17T00:00:00 | [
[
"Zhao",
"Q.",
""
],
[
"Nishida",
"T.",
""
]
] | TITLE: Using Qualitative Hypotheses to Identify Inaccurate Data
ABSTRACT: Identifying inaccurate data has long been regarded as a significant and
difficult problem in AI. In this paper, we present a new method for identifying
inaccurate data on the basis of qualitative correlations among related data.
First, we introduce the definitions of related data and qualitative
correlations among related data. Then we put forward a new concept called
support coefficient function (SCF). SCF can be used to extract, represent, and
calculate qualitative correlations among related data within a dataset. We
propose an approach to determining dynamic shift intervals of inaccurate data,
and an approach to calculating possibility of identifying inaccurate data,
respectively. Both of the approaches are based on SCF. Finally we present an
algorithm for identifying inaccurate data by using qualitative correlations
among related data as confirmatory or disconfirmatory evidence. We have
developed a practical system for interpreting infrared spectra by applying the
method, and have fully tested the system against several hundred real spectra.
The experimental results show that the method is significantly better than the
conventional methods used in many similar systems.
| no_new_dataset | 0.948585 |
1411.3374 | Manas Joglekar | Manas Joglekar, Hector Garcia-Molina, Aditya Parameswaran, Christopher
Re | Exploiting Correlations for Expensive Predicate Evaluation | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | User Defined Function(UDFs) are used increasingly to augment query languages
with extra, application dependent functionality. Selection queries involving
UDF predicates tend to be expensive, either in terms of monetary cost or
latency. In this paper, we study ways to efficiently evaluate selection queries
with UDF predicates. We provide a family of techniques for processing queries
at low cost while satisfying user-specified precision and recall constraints.
Our techniques are applicable to a variety of scenarios including when
selection probabilities of tuples are available beforehand, when this
information is available but noisy, or when no such prior information is
available. We also generalize our techniques to more complex queries. Finally,
we test our techniques on real datasets, and show that they achieve significant
savings in cost of up to $80\%$, while incurring only a small reduction in
accuracy.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 22:00:08 GMT"
}
] | 2014-11-14T00:00:00 | [
[
"Joglekar",
"Manas",
""
],
[
"Garcia-Molina",
"Hector",
""
],
[
"Parameswaran",
"Aditya",
""
],
[
"Re",
"Christopher",
""
]
] | TITLE: Exploiting Correlations for Expensive Predicate Evaluation
ABSTRACT: User Defined Function(UDFs) are used increasingly to augment query languages
with extra, application dependent functionality. Selection queries involving
UDF predicates tend to be expensive, either in terms of monetary cost or
latency. In this paper, we study ways to efficiently evaluate selection queries
with UDF predicates. We provide a family of techniques for processing queries
at low cost while satisfying user-specified precision and recall constraints.
Our techniques are applicable to a variety of scenarios including when
selection probabilities of tuples are available beforehand, when this
information is available but noisy, or when no such prior information is
available. We also generalize our techniques to more complex queries. Finally,
we test our techniques on real datasets, and show that they achieve significant
savings in cost of up to $80\%$, while incurring only a small reduction in
accuracy.
| no_new_dataset | 0.94699 |
1411.3377 | Manas Joglekar | Manas Joglekar, Hector Garcia-Molina, Aditya Parameswaran | Comprehensive and Reliable Crowd Assessment Algorithms | ICDE 2015 | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Evaluating workers is a critical aspect of any crowdsourcing system. In this
paper, we devise techniques for evaluating workers by finding confidence
intervals on their error rates. Unlike prior work, we focus on
"conciseness"---that is, giving as tight a confidence interval as possible.
Conciseness is of utmost importance because it allows us to be sure that we
have the best guarantee possible on worker error rate. Also unlike prior work,
we provide techniques that work under very general scenarios, such as when not
all workers have attempted every task (a fairly common scenario in practice),
when tasks have non-boolean responses, and when workers have different biases
for positive and negative tasks. We demonstrate conciseness as well as accuracy
of our confidence intervals by testing them on a variety of conditions and
multiple real-world datasets.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 22:09:17 GMT"
}
] | 2014-11-14T00:00:00 | [
[
"Joglekar",
"Manas",
""
],
[
"Garcia-Molina",
"Hector",
""
],
[
"Parameswaran",
"Aditya",
""
]
] | TITLE: Comprehensive and Reliable Crowd Assessment Algorithms
ABSTRACT: Evaluating workers is a critical aspect of any crowdsourcing system. In this
paper, we devise techniques for evaluating workers by finding confidence
intervals on their error rates. Unlike prior work, we focus on
"conciseness"---that is, giving as tight a confidence interval as possible.
Conciseness is of utmost importance because it allows us to be sure that we
have the best guarantee possible on worker error rate. Also unlike prior work,
we provide techniques that work under very general scenarios, such as when not
all workers have attempted every task (a fairly common scenario in practice),
when tasks have non-boolean responses, and when workers have different biases
for positive and negative tasks. We demonstrate conciseness as well as accuracy
of our confidence intervals by testing them on a variety of conditions and
multiple real-world datasets.
| no_new_dataset | 0.953188 |
1411.3409 | Paul Mineiro | Paul Mineiro, Nikos Karampatziakis | A Randomized Algorithm for CCA | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present RandomizedCCA, a randomized algorithm for computing canonical
analysis, suitable for large datasets stored either out of core or on a
distributed file system. Accurate results can be obtained in as few as two data
passes, which is relevant for distributed processing frameworks in which
iteration is expensive (e.g., Hadoop). The strategy also provides an excellent
initializer for standard iterative solutions.
| [
{
"version": "v1",
"created": "Thu, 13 Nov 2014 00:51:19 GMT"
}
] | 2014-11-14T00:00:00 | [
[
"Mineiro",
"Paul",
""
],
[
"Karampatziakis",
"Nikos",
""
]
] | TITLE: A Randomized Algorithm for CCA
ABSTRACT: We present RandomizedCCA, a randomized algorithm for computing canonical
analysis, suitable for large datasets stored either out of core or on a
distributed file system. Accurate results can be obtained in as few as two data
passes, which is relevant for distributed processing frameworks in which
iteration is expensive (e.g., Hadoop). The strategy also provides an excellent
initializer for standard iterative solutions.
| no_new_dataset | 0.94474 |
1411.3410 | Kwangchol Jang | Kwangchol Jang, Sokmin Han, Insong Kim | Person Re-identification Based on Color Histogram and Spatial
Configuration of Dominant Color Regions | 12 pages, 6 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/publicdomain/ | There is a requirement to determine whether a given person of interest has
already been observed over a network of cameras in video surveillance systems.
A human appearance obtained in one camera is usually different from the ones
obtained in another camera due to difference in illumination, pose and
viewpoint, camera parameters. Being related to appearance-based approaches for
person re-identification, we propose a novel method based on the dominant color
histogram and spatial configuration of dominant color regions on human body
parts. Dominant color histogram and spatial configuration of the dominant color
regions based on dominant color descriptor(DCD) can be considered to be robust
to illumination and pose, viewpoint changes. The proposed method is evaluated
using benchmark video datasets. Experimental results using the cumulative
matching characteristic(CMC) curve demonstrate the effectiveness of our
approach for person re-identification.
| [
{
"version": "v1",
"created": "Thu, 13 Nov 2014 00:55:48 GMT"
}
] | 2014-11-14T00:00:00 | [
[
"Jang",
"Kwangchol",
""
],
[
"Han",
"Sokmin",
""
],
[
"Kim",
"Insong",
""
]
] | TITLE: Person Re-identification Based on Color Histogram and Spatial
Configuration of Dominant Color Regions
ABSTRACT: There is a requirement to determine whether a given person of interest has
already been observed over a network of cameras in video surveillance systems.
A human appearance obtained in one camera is usually different from the ones
obtained in another camera due to difference in illumination, pose and
viewpoint, camera parameters. Being related to appearance-based approaches for
person re-identification, we propose a novel method based on the dominant color
histogram and spatial configuration of dominant color regions on human body
parts. Dominant color histogram and spatial configuration of the dominant color
regions based on dominant color descriptor(DCD) can be considered to be robust
to illumination and pose, viewpoint changes. The proposed method is evaluated
using benchmark video datasets. Experimental results using the cumulative
matching characteristic(CMC) curve demonstrate the effectiveness of our
approach for person re-identification.
| no_new_dataset | 0.951504 |
1406.2199 | Karen Simonyan | Karen Simonyan, Andrew Zisserman | Two-Stream Convolutional Networks for Action Recognition in Videos | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate architectures of discriminatively trained deep Convolutional
Networks (ConvNets) for action recognition in video. The challenge is to
capture the complementary information on appearance from still frames and
motion between frames. We also aim to generalise the best performing
hand-crafted features within a data-driven learning framework.
Our contribution is three-fold. First, we propose a two-stream ConvNet
architecture which incorporates spatial and temporal networks. Second, we
demonstrate that a ConvNet trained on multi-frame dense optical flow is able to
achieve very good performance in spite of limited training data. Finally, we
show that multi-task learning, applied to two different action classification
datasets, can be used to increase the amount of training data and improve the
performance on both.
Our architecture is trained and evaluated on the standard video actions
benchmarks of UCF-101 and HMDB-51, where it is competitive with the state of
the art. It also exceeds by a large margin previous attempts to use deep nets
for video classification.
| [
{
"version": "v1",
"created": "Mon, 9 Jun 2014 14:44:14 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Nov 2014 20:48:33 GMT"
}
] | 2014-11-13T00:00:00 | [
[
"Simonyan",
"Karen",
""
],
[
"Zisserman",
"Andrew",
""
]
] | TITLE: Two-Stream Convolutional Networks for Action Recognition in Videos
ABSTRACT: We investigate architectures of discriminatively trained deep Convolutional
Networks (ConvNets) for action recognition in video. The challenge is to
capture the complementary information on appearance from still frames and
motion between frames. We also aim to generalise the best performing
hand-crafted features within a data-driven learning framework.
Our contribution is three-fold. First, we propose a two-stream ConvNet
architecture which incorporates spatial and temporal networks. Second, we
demonstrate that a ConvNet trained on multi-frame dense optical flow is able to
achieve very good performance in spite of limited training data. Finally, we
show that multi-task learning, applied to two different action classification
datasets, can be used to increase the amount of training data and improve the
performance on both.
Our architecture is trained and evaluated on the standard video actions
benchmarks of UCF-101 and HMDB-51, where it is competitive with the state of
the art. It also exceeds by a large margin previous attempts to use deep nets
for video classification.
| no_new_dataset | 0.946101 |
1411.3041 | Ramakrishna Vedantam | Ramakrishna Vedantam, C. Lawrence Zitnick, and Devi Parikh | Collecting Image Description Datasets using Crowdsourcing | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe our two new datasets with images described by humans. Both the
datasets were collected using Amazon Mechanical Turk, a crowdsourcing platform.
The two datasets contain significantly more descriptions per image than other
existing datasets. One is based on a popular image description dataset called
the UIUC Pascal Sentence Dataset, whereas the other is based on the Abstract
Scenes dataset con- taining images made from clipart objects. In this paper we
describe our interfaces, analyze some properties of and show example
descriptions from our two datasets.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 01:34:46 GMT"
}
] | 2014-11-13T00:00:00 | [
[
"Vedantam",
"Ramakrishna",
""
],
[
"Zitnick",
"C. Lawrence",
""
],
[
"Parikh",
"Devi",
""
]
] | TITLE: Collecting Image Description Datasets using Crowdsourcing
ABSTRACT: We describe our two new datasets with images described by humans. Both the
datasets were collected using Amazon Mechanical Turk, a crowdsourcing platform.
The two datasets contain significantly more descriptions per image than other
existing datasets. One is based on a popular image description dataset called
the UIUC Pascal Sentence Dataset, whereas the other is based on the Abstract
Scenes dataset con- taining images made from clipart objects. In this paper we
describe our interfaces, analyze some properties of and show example
descriptions from our two datasets.
| new_dataset | 0.955361 |
1411.3212 | Francesco Lettich | Francesco Lettich, Salvatore Orlando, Claudio Silvestri and Christian
S. Jensen | Manycore processing of repeated range queries over massive moving
objects observations | null | null | null | null | cs.DB cs.DC cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ability to timely process significant amounts of continuously updated
spatial data is mandatory for an increasing number of applications. Parallelism
enables such applications to face this data-intensive challenge and allows the
devised systems to feature low latency and high scalability. In this paper we
focus on a specific data-intensive problem, concerning the repeated processing
of huge amounts of range queries over massive sets of moving objects, where the
spatial extents of queries and objects are continuously modified over time. To
tackle this problem and significantly accelerate query processing we devise a
hybrid CPU/GPU pipeline that compresses data output and save query processing
work. The devised system relies on an ad-hoc spatial index leading to a problem
decomposition that results in a set of independent data-parallel tasks. The
index is based on a point-region quadtree space decomposition and allows to
tackle effectively a broad range of spatial object distributions, even those
very skewed. Also, to deal with the architectural peculiarities and limitations
of the GPUs, we adopt non-trivial GPU data structures that avoid the need of
locked memory accesses and favour coalesced memory accesses, thus enhancing the
overall memory throughput. To the best of our knowledge this is the first work
that exploits GPUs to efficiently solve repeated range queries over massive
sets of continuously moving objects, characterized by highly skewed spatial
distributions. In comparison with state-of-the-art CPU-based implementations,
our method highlights significant speedups in the order of 14x-20x, depending
on the datasets, even when considering very cheap GPUs.
| [
{
"version": "v1",
"created": "Wed, 12 Nov 2014 15:46:39 GMT"
}
] | 2014-11-13T00:00:00 | [
[
"Lettich",
"Francesco",
""
],
[
"Orlando",
"Salvatore",
""
],
[
"Silvestri",
"Claudio",
""
],
[
"Jensen",
"Christian S.",
""
]
] | TITLE: Manycore processing of repeated range queries over massive moving
objects observations
ABSTRACT: The ability to timely process significant amounts of continuously updated
spatial data is mandatory for an increasing number of applications. Parallelism
enables such applications to face this data-intensive challenge and allows the
devised systems to feature low latency and high scalability. In this paper we
focus on a specific data-intensive problem, concerning the repeated processing
of huge amounts of range queries over massive sets of moving objects, where the
spatial extents of queries and objects are continuously modified over time. To
tackle this problem and significantly accelerate query processing we devise a
hybrid CPU/GPU pipeline that compresses data output and save query processing
work. The devised system relies on an ad-hoc spatial index leading to a problem
decomposition that results in a set of independent data-parallel tasks. The
index is based on a point-region quadtree space decomposition and allows to
tackle effectively a broad range of spatial object distributions, even those
very skewed. Also, to deal with the architectural peculiarities and limitations
of the GPUs, we adopt non-trivial GPU data structures that avoid the need of
locked memory accesses and favour coalesced memory accesses, thus enhancing the
overall memory throughput. To the best of our knowledge this is the first work
that exploits GPUs to efficiently solve repeated range queries over massive
sets of continuously moving objects, characterized by highly skewed spatial
distributions. In comparison with state-of-the-art CPU-based implementations,
our method highlights significant speedups in the order of 14x-20x, depending
on the datasets, even when considering very cheap GPUs.
| no_new_dataset | 0.945248 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.