id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.02785 | Charalampos Kalalas | Charalampos Kalalas, Pavol Mulinka, Guillermo Candela Belmonte, Miguel Fornell, Michail Dalgitsis, Francisco Paredes Vera, Javier Santaella S\'anchez, Carmen Vicente Villares, Roshan Sedar, Eftychia Datsika, Angelos Antonopoulos, Antonio Fern\'andez Ojea and Miquel Payaro | AI-Driven Vehicle Condition Monitoring with Cell-Aware Edge Service Migration | 6 pages, 8 figures | null | null | null | cs.NI cs.AI | http://creativecommons.org/licenses/by/4.0/ | Artificial intelligence (AI) has been increasingly applied to the condition monitoring of vehicular equipment, aiming to enhance maintenance strategies, reduce costs, and improve safety. Leveraging the edge computing paradigm, AI-based condition monitoring systems process vast streams of vehicular data to detect anomalies and optimize operational performance. In this work, we introduce a novel vehicle condition monitoring service that enables real-time diagnostics of a diverse set of anomalies while remaining practical for deployment in real-world edge environments. To address mobility challenges, we propose a closed-loop service orchestration framework where service migration across edge nodes is dynamically triggered by network-related metrics. Our approach has been implemented and tested in a real-world race circuit environment equipped with 5G network capabilities under diverse operational conditions. Experimental results demonstrate the effectiveness of our framework in ensuring low-latency AI inference and adaptive service placement, highlighting its potential for intelligent transportation and mobility applications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:12:27 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kalalas",
"Charalampos",
""
],
[
"Mulinka",
"Pavol",
""
],
[
"Belmonte",
"Guillermo Candela",
""
],
[
"Fornell",
"Miguel",
""
],
[
"Dalgitsis",
"Michail",
""
],
[
"Vera",
"Francisco Paredes",
""
],
[
"Sánchez",
"Javier Santaella",
""
],
[
"Villares",
"Carmen Vicente",
""
],
[
"Sedar",
"Roshan",
""
],
[
"Datsika",
"Eftychia",
""
],
[
"Antonopoulos",
"Angelos",
""
],
[
"Ojea",
"Antonio Fernández",
""
],
[
"Payaro",
"Miquel",
""
]
] |
2506.02786 | Haitang Yang | Xin Jiang and Haitang Yang | Entanglement Entropy of Conformal Field Theory in All Dimensions | 14 pages, 6 figures | null | null | CTP-SCU/2025011 | hep-th gr-qc | http://creativecommons.org/licenses/by/4.0/ | We provide a field-theoretic method to calculate entanglement entropy of CFT in all dimensions, for arbitrary shape of entangling surfaces. Ryu-Takayanagi formula is manifested in this formalism. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:12:31 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Jiang",
"Xin",
""
],
[
"Yang",
"Haitang",
""
]
] |
2506.02787 | Ruilong Wu | Ruilong Wu, Xinjiao Li, Yisu Wang, Xinyu Chen, Dirk Kutscher | Rethinking Dynamic Networks and Heterogeneous Computing with Automatic Parallelization | null | null | null | null | cs.DC cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hybrid parallelism techniques are essential for efficiently training large language models (LLMs). Nevertheless, current automatic parallel planning frameworks often overlook the simultaneous consideration of node heterogeneity and dynamic network topology changes, limiting their effectiveness in practical applications. In this paper, we address these limitations by modeling heterogeneous nodes within dynamically changing network environments and leveraging simulation-based strategies to determine optimal parallel configurations. Our approach enables fine-grained workload allocation tailored for heterogeneous nodes and complex network scenarios, achieving performance competitive with state-of-the-art methods under regular and stable network conditions. Additionally, we introduce a strategy pruning technique to rapidly discard infeasible parallel configurations, substantially reducing the search space and accelerating the search process through parallel execution within the simulator. Preliminary evaluations confirm that our method notably enhances training performance on heterogeneous nodes and demonstrates improved adaptability in complex, dynamic scenarios such as cloud computing environments. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:14:17 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wu",
"Ruilong",
""
],
[
"Li",
"Xinjiao",
""
],
[
"Wang",
"Yisu",
""
],
[
"Chen",
"Xinyu",
""
],
[
"Kutscher",
"Dirk",
""
]
] |
2506.02788 | Muhammad Shamrooz Aslam Prof. | Muhammad Shamrooz Aslam, Hazrat Bilal, Sumeera Shamrooz | Quantized Dissipative Uncertain Model for Fractional T_S Fuzzy systems with Time_Varying Delays Under Networked Control System | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper addressed with the quantized dissipative uncertain problem for delayed fractional T_S Fuzzy system for event_triggered networked systems (E_NS), where the extended dissipativity analysis combines the H infinity, dissipativity, L2 and L infinity and passivity performance in a unified frame. To attain the high efficiency for available channel resources, measurement size decrease mechanism and event_triggered scheme (ETS) are proposed. Firstly, we present the ETS in which signal is transmitted through the channel with logical function then logarithmic quantization methodology is implemented for size reduction. Then, we transfer the original delayed fractional T_S fuzzy systems with the effect of quantization under ETS as induced communications delays. Furthermore, by employing the associative Lyapunov functional method in terms of linear matrix inequalities, adequate conditions for asymptotical stability is given. Moreover, we also construct the design fuzzy model for state space filtering system. At last, a truck_trailer model is given to show the effectiveness of the proposed strategy. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:14:19 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Aslam",
"Muhammad Shamrooz",
""
],
[
"Bilal",
"Hazrat",
""
],
[
"Shamrooz",
"Sumeera",
""
]
] |
2506.02789 | Peiqi Li | Renxing Li, Weiyi Tang, Peiqi Li, Qiming Huang, Jiayuan She, Shengkai Li, Haoran Xu, Yeyun Wan, Jing Liu, Hailong Fu, Xiang Li, Jiangang Chen | Automated Measurement of Optic Nerve Sheath Diameter Using Ocular Ultrasound Video | 17 pages, 9 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Objective. Elevated intracranial pressure (ICP) is recognized as a biomarker of secondary brain injury, with a significant linear correlation observed between optic nerve sheath diameter (ONSD) and ICP. Frequent monitoring of ONSD could effectively support dynamic evaluation of ICP. However, ONSD measurement is heavily reliant on the operator's experience and skill, particularly in manually selecting the optimal frame from ultrasound sequences and measuring ONSD. Approach. This paper presents a novel method to automatically identify the optimal frame from video sequences for ONSD measurement by employing the Kernel Correlation Filter (KCF) tracking algorithm and Simple Linear Iterative Clustering (SLIC) segmentation algorithm. The optic nerve sheath is mapped and measured using a Gaussian Mixture Model (GMM) combined with a KL-divergence-based method. Results. When compared with the average measurements of two expert clinicians, the proposed method achieved a mean error, mean squared deviation, and intraclass correlation coefficient (ICC) of 0.04, 0.054, and 0.782, respectively. Significance. The findings suggest that this method provides highly accurate automated ONSD measurements, showing potential for clinical application. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:14:51 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Li",
"Renxing",
""
],
[
"Tang",
"Weiyi",
""
],
[
"Li",
"Peiqi",
""
],
[
"Huang",
"Qiming",
""
],
[
"She",
"Jiayuan",
""
],
[
"Li",
"Shengkai",
""
],
[
"Xu",
"Haoran",
""
],
[
"Wan",
"Yeyun",
""
],
[
"Liu",
"Jing",
""
],
[
"Fu",
"Hailong",
""
],
[
"Li",
"Xiang",
""
],
[
"Chen",
"Jiangang",
""
]
] |
2506.02790 | Shunxin Yao | Shunxin Yao | Orthogonality-Constrained Deep Instrumental Variable Model for Causal Effect Estimation | 19 pages, 3 figures, 1 table | null | null | null | econ.EM | http://creativecommons.org/publicdomain/zero/1.0/ | OC-DeepIV is a neural network model designed for estimating causal effects. It characterizes heterogeneity by adding interaction features and reduces redundancy through orthogonal constraints. The model includes two feature extractors, one for the instrumental variable Z and the other for the covariate X*. The training process is divided into two stages: the first stage uses the mean squared error (MSE) loss function, and the second stage incorporates orthogonal regularization. Experimental results show that this model outperforms DeepIV and DML in terms of accuracy and stability. Future research directions include applying the model to real-world problems and handling scenarios with multiple processing variables. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:15:03 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Yao",
"Shunxin",
""
]
] |
2506.02791 | Hongyi Lin | Zhen Yang, Hongyi Lin, Yifan He, Jie Xu, Zeyu Sun, Shuo Liu, Pengpeng Wang, Zhongxing Yu, Qingyuan Liang | Rethinking the effects of data contamination in Code Intelligence | null | null | null | null | cs.SE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, code intelligence has gained increasing importance in the field of automated software engineering. Meanwhile, the widespread adoption of Pretrained Language Models (PLMs) and Large Language Models (LLMs) has raised concerns regarding data contamination and its potential impact on model performance evaluation. This paper presents a systematic empirical study to investigate the fine-grained data contamination on code intelligence tasks. Our study involves diverse representative PLMs, namely RoBERTa and GPT-2, and LLMs, namely LLaMA and StarCoder, covering three major tasks: code translation, code generation, and code summarization. We categorize contamination scenarios into four types according to the code intelligence practice, namely input-only, output-only, unpaired, and paired contamination settings, and construct corresponding experimental and control groups for exploration.
Experimental results show that, under the pre-training, fine-tuning, and inference paradigm adopted by PLMs, even deliberately injecting paired contamination does not lead to significant performance overestimation. But direct inference or small-scale fine-tuning uncovers the contamination effects. In contrast, LLMs with pre-training and inference paradigm are significantly affected by the paired contamination. Apart from the above, other contamination scenarios have no impact on both PLMs and LLMs. Our findings challenge the conventional belief that contamination inevitably leads to performance overestimation, providing new insights into the evaluation and deployment of code intelligence models. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:15:44 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Yang",
"Zhen",
""
],
[
"Lin",
"Hongyi",
""
],
[
"He",
"Yifan",
""
],
[
"Xu",
"Jie",
""
],
[
"Sun",
"Zeyu",
""
],
[
"Liu",
"Shuo",
""
],
[
"Wang",
"Pengpeng",
""
],
[
"Yu",
"Zhongxing",
""
],
[
"Liang",
"Qingyuan",
""
]
] |
2506.02792 | Ayesha Afzal | Ayesha Afzal, Georg Hager, Gerhard Wellen | Exploring metrics for analyzing dynamic behavior in MPI programs via a coupled-oscillator model | null | null | null | null | cs.DC physics.app-ph physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel, lightweight, and physically inspired approach to modeling the dynamics of parallel distributed-memory programs. Inspired by the Kuramoto model, we represent MPI processes as coupled oscillators with topology-aware interactions, custom coupling potentials, and stochastic noise. The resulting system of nonlinear ordinary differential equations opens a path to modeling key performance phenomena of parallel programs, including synchronization, delay propagation and decay, bottlenecks, and self-desynchronization.
This paper introduces interaction potentials to describe memory- and compute-bound workloads and employs multiple quantitative metrics -- such as an order parameter, synchronization entropy, phase gradients, and phase differences -- to evaluate phase coherence and disruption. We also investigate the role of local noise and show that moderate noise can accelerate resynchronization in scalable applications. Our simulations align qualitatively with MPI trace data, showing the potential of physics-informed abstractions to predict performance patterns, which offers a new perspective for performance modeling and software-hardware co-design in parallel computing. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:15:49 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Afzal",
"Ayesha",
""
],
[
"Hager",
"Georg",
""
],
[
"Wellen",
"Gerhard",
""
]
] |
2506.02793 | Houssam Zenati | Houssam Zenati, Bariscan Bozkurt, Arthur Gretton | Doubly-Robust Estimation of Counterfactual Policy Mean Embeddings | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Estimating the distribution of outcomes under counterfactual policies is critical for decision-making in domains such as recommendation, advertising, and healthcare. We analyze a novel framework-Counterfactual Policy Mean Embedding (CPME)-that represents the entire counterfactual outcome distribution in a reproducing kernel Hilbert space (RKHS), enabling flexible and nonparametric distributional off-policy evaluation. We introduce both a plug-in estimator and a doubly robust estimator; the latter enjoys improved uniform convergence rates by correcting for bias in both the outcome embedding and propensity models. Building on this, we develop a doubly robust kernel test statistic for hypothesis testing, which achieves asymptotic normality and thus enables computationally efficient testing and straightforward construction of confidence intervals. Our framework also supports sampling from the counterfactual distribution. Numerical simulations illustrate the practical benefits of CPME over existing methods. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:16:46 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Zenati",
"Houssam",
""
],
[
"Bozkurt",
"Bariscan",
""
],
[
"Gretton",
"Arthur",
""
]
] |
2506.02794 | Mijeong Kim | Mijeong Kim, Gunhee Kim, Jungyoon Choi, Wonjae Roh, Bohyung Han | PhysGaia: A Physics-Aware Dataset of Multi-Body Interactions for Dynamic Novel View Synthesis | Project page: http://cvlab.snu.ac.kr/research/PhysGaia, Data: https://huggingface.co/datasets/mijeongkim/PhysGaia/tree/main | null | null | null | cs.GR cs.AI cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We introduce PhysGaia, a novel physics-aware dataset specifically designed for Dynamic Novel View Synthesis (DyNVS), encompassing both structured objects and unstructured physical phenomena. Unlike existing datasets that primarily focus on photorealistic reconstruction, PhysGaia is created to actively support physics-aware dynamic scene modeling. Our dataset provides complex dynamic scenarios with rich interactions among multiple objects, where they realistically collide with each other and exchange forces. Furthermore, it contains a diverse range of physical materials, such as liquid, gas, viscoelastic substance, and textile, which moves beyond the rigid bodies prevalent in existing datasets. All scenes in PhysGaia are faithfully generated to strictly adhere to physical laws, leveraging carefully selected material-specific physics solvers. To enable quantitative evaluation of physical modeling, our dataset provides essential ground-truth information, including 3D particle trajectories and physics parameters, e.g., viscosity. To facilitate research adoption, we also provide essential integration pipelines for using state-of-the-art DyNVS models with our dataset and report their results. By addressing the critical lack of datasets for physics-aware modeling, PhysGaia will significantly advance research in dynamic view synthesis, physics-based scene understanding, and deep learning models integrated with physical simulation -- ultimately enabling more faithful reconstruction and interpretation of complex dynamic scenes. Our datasets and codes are available in the project website, http://cvlab.snu.ac.kr/research/PhysGaia. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:19:18 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kim",
"Mijeong",
""
],
[
"Kim",
"Gunhee",
""
],
[
"Choi",
"Jungyoon",
""
],
[
"Roh",
"Wonjae",
""
],
[
"Han",
"Bohyung",
""
]
] |
2506.02795 | Guangqi Wu | Guangqi Wu, Runzhong Wang, Connor W. Coley | Optimization of Robotic Liquid Handling as a Capacitated Vehicle Routing Problem | null | null | null | null | math.OC cs.RO | http://creativecommons.org/licenses/by/4.0/ | We present an optimization strategy to reduce the execution time of liquid handling operations in the context of an automated chemical laboratory. By formulating the task as a capacitated vehicle routing problem (CVRP), we leverage heuristic solvers traditionally used in logistics and transportation planning to optimize task execution times. As exemplified using an 8-channel pipette with individually controllable tips, our approach demonstrates robust optimization performance across different labware formats (e.g., well-plates, vial holders), achieving up to a 37% reduction in execution time for randomly generated tasks compared to the baseline sorting method. We further apply the method to a real-world high-throughput materials discovery campaign and observe that 3 minutes of optimization time led to a reduction of 61 minutes in execution time compared to the best-performing sorting-based strategy. Our results highlight the potential for substantial improvements in throughput and efficiency in automated laboratories without any hardware modifications. This optimization strategy offers a practical and scalable solution to accelerate combinatorial experimentation in areas such as drug combination screening, reaction condition optimization, materials development, and formulation engineering. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:22:33 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wu",
"Guangqi",
""
],
[
"Wang",
"Runzhong",
""
],
[
"Coley",
"Connor W.",
""
]
] |
2506.02796 | Chen Liu | Haoyuan Wang, Chen Liu, Minh-Ngoc Tran, Chao Wang | Deep Learning Enhanced Multivariate GARCH | null | null | null | null | q-fin.CP cs.AI econ.EM | http://creativecommons.org/licenses/by-sa/4.0/ | This paper introduces a novel multivariate volatility modeling framework, named Long Short-Term Memory enhanced BEKK (LSTM-BEKK), that integrates deep learning into multivariate GARCH processes. By combining the flexibility of recurrent neural networks with the econometric structure of BEKK models, our approach is designed to better capture nonlinear, dynamic, and high-dimensional dependence structures in financial return data. The proposed model addresses key limitations of traditional multivariate GARCH-based methods, particularly in capturing persistent volatility clustering and asymmetric co-movement across assets. Leveraging the data-driven nature of LSTMs, the framework adapts effectively to time-varying market conditions, offering improved robustness and forecasting performance. Empirical results across multiple equity markets confirm that the LSTM-BEKK model achieves superior performance in terms of out-of-sample portfolio risk forecast, while maintaining the interpretability from the BEKK models. These findings highlight the potential of hybrid econometric-deep learning models in advancing financial risk management and multivariate volatility forecasting. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:22:57 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wang",
"Haoyuan",
""
],
[
"Liu",
"Chen",
""
],
[
"Tran",
"Minh-Ngoc",
""
],
[
"Wang",
"Chao",
""
]
] |
2506.02797 | Paul Didier | Paul Didier, Toon van Waterschoot, Simon Doclo, J\"org Bitzer, Marc Moonen | Fast-Converging Distributed Signal Estimation in Topology-Unconstrained Wireless Acoustic Sensor Networks | null | null | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by/4.0/ | This paper focuses on distributed signal estimation in topology-unconstrained wireless acoustic sensor networks (WASNs) where sensor nodes only transmit fused versions of their local sensor signals. For this task, the topology-independent (TI) distributed adaptive node-specific signal estimation (DANSE) algorithm (TI-DANSE) has previously been proposed. It converges towards the centralized signal estimation solution in non-fully connected and time-varying network topologies. However, the applicability of TI-DANSE in real-world scenarios is limited due to its slow convergence. The latter results from the fact that, in TI-DANSE, nodes only have access to the in-network sum of all fused signals in the WASN. We address this low convergence speed by introducing an improved TI-DANSE algorithm, referred to as TI-DANSE+, in which updating nodes separately use the partial in-network sums of fused signals coming from each of their neighbors. Nodes can maximize the number of available degrees of freedom in their local optimization problem, leading to faster convergence. This is further exploited by combining TI-DANSE+ with a tree-pruning strategy that maximizes the number of neighbors at the updating node. In fully connected WASNs, TI-DANSE+ converges as fast as the original DANSE algorithm (the latter only defined for fully connected WASNs) while using peer-to-peer data transmission instead of broadcasting and thus saving communication bandwidth. If link failures occur, the convergence of TI-DANSE+ towards the centralized solution is preserved without any change in its formulation. Altogether, the proposed TI-DANSE+ algorithm can be viewed as an all-round alternative to DANSE and TI-DANSE which (i) merges the advantages of both, (ii) reconciliates their differences into a single formulation, and (iii) shows advantages of its own in terms of communication bandwidth usage. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:23:18 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Didier",
"Paul",
""
],
[
"van Waterschoot",
"Toon",
""
],
[
"Doclo",
"Simon",
""
],
[
"Bitzer",
"Jörg",
""
],
[
"Moonen",
"Marc",
""
]
] |
2506.02798 | Holly Huddleston | H.M. Huddleston, M. Yeung, C.R.J. Fitzpatrick, J.P. Kennedy, S. Palaniyppan, R. Shah, D. C. Gautier, M. Zepf, J. C. Fernandez, B. M. Hegelich, and B. Dromey | Diagnosing the impact of relativistically intense prepulse on few-picosecond timeframes for short scale length laser-matter interactions | 7 pages, 3 figures | null | null | null | physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid proliferation of multi-petawatt (MPW) lasers globally, a new era of high-energy density science promises to emerge within the next decade. However, precise control over how light at these ultra-relativistic intensities interacts with matter (especially with solid-density targets) will be crucial to fully realize the cutting-edge scientific advancements and technological breakthroughs that the MPW regime promises to unlock. In this manuscript, we present experimental results, supported by numerical simulations, which show how intense prepulse activity on few-ps ($10^{-12}$ s) timescales leads to rapid shifts in the steepness of the preplasma generated on the surface of ultra-thin nanofoil targets. By combining a single-shot frequency resolved optical gating (FROG) autocorrelation device to diagnose on-shot incident laser pulse contrast, with coherent synchrotron emission (CSE) from relativistic laser plasmas as a probe for evolving plasma-scale length conditions, we provide an experimental benchmark for laser contrast on forthcoming MPW facilities, where high contrast on few-ps timescales will be essential for the next generation of laser-solid interactions. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:24:22 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Huddleston",
"H. M.",
""
],
[
"Yeung",
"M.",
""
],
[
"Fitzpatrick",
"C. R. J.",
""
],
[
"Kennedy",
"J. P.",
""
],
[
"Palaniyppan",
"S.",
""
],
[
"Shah",
"R.",
""
],
[
"Gautier",
"D. C.",
""
],
[
"Zepf",
"M.",
""
],
[
"Fernandez",
"J. C.",
""
],
[
"Hegelich",
"B. M.",
""
],
[
"Dromey",
"B.",
""
]
] |
2506.02799 | Debangan Maji | Debangan Maji, Subhashis Roy | Distance to the Galactic centre 7$^\prime$ halo from HI absorption | Accepted for publication in ApJ | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Using archival data, we have made an HI absorption study of the 7$^\prime$ halo surrounding the Sgr A complex, observed towards the Galactic centre (GC) region. We find strong HI absorption near velocities of $-$53 km s$^{-1}$, which is due to the 3-kpc arm, placing it beyond 5 kpc from us. We further examined the HI absorption properties towards 5 different parts of the 7$^\prime$ halo. Absorption by +50 km s$^{-1}$ GC cloud is seen towards only 3 parts of the halo, but not towards the other 2 regions. Strong emissions in CO and CS are, however, identified toward all the above 5 parts of the halo by the +50 km s$^{-1}$ GC molecular cloud. This does show that the 7$^\prime$ halo is partly behind, and partly in front of the +50 km s$^{-1}$ cloud. To our knowledge, this, for the first time clearly shows the 7$^\prime$ halo to be located at the same distance as the +50 km s$^{-1}$ molecular cloud, i.e., at the GC region. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:29:50 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Maji",
"Debangan",
""
],
[
"Roy",
"Subhashis",
""
]
] |
2506.02800 | Jie Guo | Jie Guo, Quansen Jiu | On Stability and Instability of the De Gregorio Model Near Excited States on the Torus | null | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | The Constantin-Lax-Majda (CLM) model and the De Gregorio model which is a modification of the CLM model are well-known for their ability to emulate the behavior of the 3D Euler equations, particularly their potential to develop finite-time singularities. The stability properties of the De Gregorio model on the torus near the ground state $-\sin\theta$ have been thoroughly studied. However, the stability analysis near excited states $-\sin k\theta$ with $k\ge 2$ remains challenging. This paper focuses on analyzing the stability and instability of the De Gregorio model on torus around the first excited state $-\sin 2\theta$. The linear and nonlinear instability are established for a broad class of initial data, while nonlinear stability is proved for another large class of initial data in this paper. Our analysis reveals that solution behavior to the De Gregorio model near excited states demonstrates different stability patterns depending on initial conditions. One of new ingredients in our instability analysis involves deriving a second-order ordinary differential equation (ODE) governing the Fourier coefficients of solutions and examining the spectral properties of a positive definite quadratic form emerging from this ODE. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:31:20 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Guo",
"Jie",
""
],
[
"Jiu",
"Quansen",
""
]
] |
2506.02801 | Juan Carlos Buitrago Oropeza | J.C. Buitrago Oropeza | Concentration of the maximum size of an induced subtree in moderately sparse random graphs | 27 pages | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Kamaldinov, Skorkin, and Zhukovskii proved that the maximum size of an induced subtree in the binomial random graph $G(n,p)$ is concentrated at two consecutive points, whenever $p\in(0,1)$ is a constant. Using improved bounds on the second moment of the number of induced subtrees, we show that the same result holds when $n^{-\frac{e-2}{3e-2}+\varepsilon}\leq p=o(1)$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:32:26 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Oropeza",
"J. C. Buitrago",
""
]
] |
2506.02802 | Ioana Giurgiu | Andr\'as Strausz, Niels Pardon, Ioana Giurgiu | A Learned Cost Model-based Cross-engine Optimizer for SQL Workloads | 6 pages | null | null | null | cs.DB cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Lakehouse systems enable the same data to be queried with multiple execution engines. However, selecting the engine best suited to run a SQL query still requires a priori knowledge of the query computational requirements and an engine capability, a complex and manual task that only becomes more difficult with the emergence of new engines and workloads. In this paper, we address this limitation by proposing a cross-engine optimizer that can automate engine selection for diverse SQL queries through a learned cost model. Optimized with hints, a query plan is used for query cost prediction and routing. Cost prediction is formulated as a multi-task learning problem, and multiple predictor heads, corresponding to different engines and provisionings, are used in the model architecture. This eliminates the need to train engine-specific models and allows the flexible addition of new engines at a minimal fine-tuning cost. Results on various databases and engines show that using a query optimized logical plan for cost estimation decreases the average Q-error by even 12.6% over using unoptimized plans as input. Moreover, the proposed cross-engine optimizer reduces the total workload runtime by up to 25.2% in a zero-shot setting and 30.4% in a few-shot setting when compared to random routing. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:32:56 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Strausz",
"András",
""
],
[
"Pardon",
"Niels",
""
],
[
"Giurgiu",
"Ioana",
""
]
] |
2506.02803 | Sifan Li | Sifan Li and Yujun Cai and Yiwei Wang | SemVink: Advancing VLMs' Semantic Understanding of Optical Illusions via Visual Global Thinking | null | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-language models (VLMs) excel in semantic tasks but falter at a core human capability: detecting hidden content in optical illusions or AI-generated images through perceptual adjustments like zooming. We introduce HC-Bench, a benchmark of 112 images with hidden text, objects, and illusions, revealing that leading VLMs achieve near-zero accuracy (0-5.36%)-even with explicit prompting. Humans resolve such ambiguities instinctively, yet VLMs fail due to an overreliance on high-level semantics. Strikingly, we propose SemVink (Semantic Visual Thinking) by simply scaling images to low resolutions (32-128 pixels), which unlocks >99% accuracy by eliminating redundant visual noise. This exposes a critical architectural flaw: VLMs prioritize abstract reasoning over low-level visual operations crucial for real-world robustness. Our work urges a shift toward hybrid models integrating multi-scale processing, bridging the gap between computational vision and human cognition for applications in medical imaging, security, and beyond. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:33:47 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Li",
"Sifan",
""
],
[
"Cai",
"Yujun",
""
],
[
"Wang",
"Yiwei",
""
]
] |
2506.02804 | Megan Schwamb | Stephanie R. Merritt, Grigori Fedorets, Megan E. Schwamb, Samuel Cornwall, Pedro H. Bernardinelli, Mario Juric, Matthew J. Holman, Jacob A. Kurlander, Siegfried Eggl, Drew Oldag, Maxine West, Jeremy Kubica, Joseph Murtagh, R. Lynne Jones, Peter Yoachim, Ryan R. Lyttle, Michael S. P. Kelley, Joachim Moeyens, Kathleen Kiker, Shantanu P. Naidu, Colin Snodgrass, Shannon M. Matthews, and Colin Orion Chandler | Sorcha: A Solar System Survey Simulator for the Legacy Survey of Space and Time | Accepted to AJ, 64 pages | null | null | null | astro-ph.EP astro-ph.IM | http://creativecommons.org/licenses/by/4.0/ | The upcoming Legacy Survey of Space and Time (LSST) at the Vera C. Rubin Observatory is expected to revolutionize solar system astronomy. Unprecedented in scale, this ten-year wide-field survey will collect billions of observations and discover a predicted $\sim$5 million new solar system objects. Like all astronomical surveys, its results will be affected by a complex system of intertwined detection biases. Survey simulators have long been used to forward-model the effects of these biases on a given population, allowing for a direct comparison to real discoveries. However, the scale and tremendous scope of the LSST requires the development of new tools. In this paper we present Sorcha, an open-source survey simulator written in Python. Designed with the scale of LSST in mind, Sorcha is a comprehensive survey simulator to cover all solar system small-body populations. Its flexible, modular design allows Sorcha to be easily adapted to other surveys by the user. The simulator is built to run both locally and on high-performance computing (HPC) clusters, allowing for repeated simulation of millions to billions of objects (both real and synthetic). | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:34:32 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Merritt",
"Stephanie R.",
""
],
[
"Fedorets",
"Grigori",
""
],
[
"Schwamb",
"Megan E.",
""
],
[
"Cornwall",
"Samuel",
""
],
[
"Bernardinelli",
"Pedro H.",
""
],
[
"Juric",
"Mario",
""
],
[
"Holman",
"Matthew J.",
""
],
[
"Kurlander",
"Jacob A.",
""
],
[
"Eggl",
"Siegfried",
""
],
[
"Oldag",
"Drew",
""
],
[
"West",
"Maxine",
""
],
[
"Kubica",
"Jeremy",
""
],
[
"Murtagh",
"Joseph",
""
],
[
"Jones",
"R. Lynne",
""
],
[
"Yoachim",
"Peter",
""
],
[
"Lyttle",
"Ryan R.",
""
],
[
"Kelley",
"Michael S. P.",
""
],
[
"Moeyens",
"Joachim",
""
],
[
"Kiker",
"Kathleen",
""
],
[
"Naidu",
"Shantanu P.",
""
],
[
"Snodgrass",
"Colin",
""
],
[
"Matthews",
"Shannon M.",
""
],
[
"Chandler",
"Colin Orion",
""
]
] |
2506.02805 | Henri Bollaert | Henri Bollaert, Chris Cornelis, Marko Palangeti\'c, Salvatore Greco, Roman S{\l}owi\'nski | Optimising the attribute order in Fuzzy Rough Rule Induction | This is the author's version of the work accepted for publication in Lecture Notes in Computer Science. The final publication is available at Springer via https://doi.org/10.1007/978-3-031-92747-8_16 | null | 10.1007/978-3-031-92747-8_16 | null | cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Interpretability is the next pivotal frontier in machine learning research. In the pursuit of glass box models - as opposed to black box models, like random forests or neural networks - rule induction algorithms are a logical and promising avenue, as the rules can easily be understood by humans. In our previous work, we introduced FRRI, a novel rule induction algorithm based on fuzzy rough set theory. We demonstrated experimentally that FRRI outperformed other rule induction methods with regards to accuracy and number of rules. FRRI leverages a fuzzy indiscernibility relation to partition the data space into fuzzy granules, which are then combined into a minimal covering set of rules. This indiscernibility relation is constructed by removing attributes from rules in a greedy way. This raises the question: does the order of the attributes matter? In this paper, we show that optimising only the order of attributes using known methods from fuzzy rough set theory and classical machine learning does not improve the performance of FRRI on multiple metrics. However, removing a small number of attributes using fuzzy rough feature selection during this step positively affects balanced accuracy and the average rule length. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:34:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Bollaert",
"Henri",
""
],
[
"Cornelis",
"Chris",
""
],
[
"Palangetić",
"Marko",
""
],
[
"Greco",
"Salvatore",
""
],
[
"Słowiński",
"Roman",
""
]
] |
2506.02806 | Isabella Ianni | Abdelrazek Dieb, Isabella Ianni | A note on a Pohozaev identity for the fractional Green function | null | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We get a Pohozaev-type identity for the fractional Green function, which extends to the fractional setting a classical result by Brezis and Peletier. Our result complements with some more recent ones obtained by Djitte and Sueur concerning a representation formula for the gradient of the fractional Robin function. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:34:57 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Dieb",
"Abdelrazek",
""
],
[
"Ianni",
"Isabella",
""
]
] |
2506.02807 | Cesar Pairetti | Fabrizio E. Aguzzi, Santiago M. Rabazzi, Mart\'in S. Armoa, C\'esar I. Pairetti and Alejandro E. Albanesi | An open-source finite element toolbox for anisotropic creep and irradiation growth: Application to tube and spacer grid assembly | preprint on review | null | null | null | physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | This work presents an open-source interface that couples the viscoplastic self-consistent (VPSC) model capable of simulating anisotropic creep and irradiation growth in polycrystalline materials with the finite element solver Code_Aster. The interface enables the simulation of the micromechanical response of irradiated zirconium alloy components by integrating grain-level constitutive behavior into a structural FEM framework. A key feature is the automated rotation of stress and strain tensors between the global FEM frame and the local crystallographic axes, a transformation not natively supported by Code_Aster. The elastic strain is recovered analytically using the inverse of the self-consistently stiffness tensor provided by VPSC. As a demonstration, the framework is applied to an actual model of a pressurized water reactor (PWR) spacer grid, based on a patented design, capturing nonlinear contact and the anisotropic response of the cladding and grid. Simulations reveal the micromechanisms controlling the evolution of clearance between components and highlight the role of crystallographic texture in mitigating wear. In particular, a texture with a high fraction of prismatic planes oriented in the normal direction of the grid appears to be the most suitable for spacer design, as it minimizes clearance and contributes to wear resistance. The interface offers a flexible, extensible platform for high-fidelity simulations in nuclear fuel performance analysis. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:35:37 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Aguzzi",
"Fabrizio E.",
""
],
[
"Rabazzi",
"Santiago M.",
""
],
[
"Armoa",
"Martín S.",
""
],
[
"Pairetti",
"César I.",
""
],
[
"Albanesi",
"Alejandro E.",
""
]
] |
2506.02808 | Christian Meyer | Christian Meyer, Gerd Wachsmuth | Optimal control of the Poisson equation with transport regularization: Properties of optimal transport plans and transport map | null | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An optimal control problem in the space of Borel measures governed by the Poisson equation is investigated. The characteristic feature of the problem under consideration is the Tikhonov regularization term in form of the transportation distance of the control to a given prior. Existence of optimal solutions is shown and first-order necessary optimality conditions are derived. The latter are used to deduce structural a priori information about the optimal control and its support based on properties of the associated optimal transport plan. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:36:36 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Meyer",
"Christian",
""
],
[
"Wachsmuth",
"Gerd",
""
]
] |
2506.02809 | Reyhaneh Khasseh | M. A. Rajabpour, MirAdel Seifi MirJafarlou, and Reyhaneh Khasseh | Matrix Elements of Fermionic Gaussian Operators in Arbitrary Pauli Bases: A Pfaffian Formula | 27 pages | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fermionic Gaussian operators are foundational tools in quantum many-body theory, numerical simulation of fermionic dynamics, and fermionic linear optics. While their structure is fully determined by two-point correlations, evaluating their matrix elements in arbitrary local spin bases remains a nontrivial task, especially in applications involving quantum measurements, tomography, and basis-rotated simulations. In this work, we derive a fully explicit and general Pfaffian formula for the matrix elements of fermionic Gaussian operators between arbitrary Pauli product states. Our approach introduces a pair of sign-encoding matrices whose classification leads to a Lie algebra isomorphic to $\mathfrak{so}(2L)$. This algebraic structure not only guarantees consistency of the Pfaffian signs but also reveals deep connections to Clifford algebras. The resulting framework enables scalable computations across diverse fields -- from quantum tomography and entanglement dynamics to algebraic structure in fermionic circuits and matchgate computation. Beyond its practical utility, our construction sheds light on the internal symmetries of Gaussian operators and offers a new lens through which to explore their role in quantum information and computational models. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:37:06 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Rajabpour",
"M. A.",
""
],
[
"MirJafarlou",
"MirAdel Seifi",
""
],
[
"Khasseh",
"Reyhaneh",
""
]
] |
2506.02810 | Ziyad Oulhaj | Ziyad Oulhaj, Mathieu Carri\`ere and Bertrand Michel | Gromov-Wasserstein Bound between Reeb and Mapper Graphs | null | null | null | null | math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | Since its introduction as a computable approximation of the Reeb graph, the Mapper graph has become one of the most popular tools from topological data analysis for performing data visualization and inference. However, finding an appropriate metric (that is, a tractable metric with theoretical guarantees) for comparing Reeb and Mapper graphs, in order to, e.g., quantify the rate of convergence of the Mapper graph to the Reeb graph, is a difficult problem. While several metrics have been proposed in the literature, none is able to incorporate measure information, when data points are sampled according to an underlying probability measure. The resulting Reeb and Mapper graphs are therefore purely deterministic and combinatorial, and substantial effort is thus required to ensure their statistical validity. In this article, we handle this issue by treating Reeb and Mapper graphs as metric measure spaces. This allows us to use Gromov-Wasserstein metrics to compare these graphs directly in order to better incorporate the probability measures that data points are sampled from. Then, we describe the geometry that arises from this perspective, and we derive rates of convergence of the Mapper graph to the Reeb graph in this context. Finally, we showcase the usefulness of such metrics for Reeb and Mapper graphs in a few numerical experiments. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:37:19 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Oulhaj",
"Ziyad",
""
],
[
"Carrière",
"Mathieu",
""
],
[
"Michel",
"Bertrand",
""
]
] |
2506.02811 | Ant\'onio Pedro Correia Da Silva Pinheiro | Ant\'onio Pedro Pinheiro, Rita P. Ribeiro | CART-based Synthetic Tabular Data Generation for Imbalanced Regression | 15 pages, 2 figures, 5 tables, 1 algorithm | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Handling imbalanced target distributions in regression tasks remains a significant challenge in tabular data settings where underrepresented regions can hinder model performance. Among data-level solutions, some proposals, such as random sampling and SMOTE-based approaches, propose adapting classification techniques to regression tasks. However, these methods typically rely on crisp, artificial thresholds over the target variable, a limitation inherited from classification settings that can introduce arbitrariness, often leading to non-intuitive and potentially misleading problem formulations. While recent generative models, such as GANs and VAEs, provide flexible sample synthesis, they come with high computational costs and limited interpretability. In this study, we propose adapting an existing CART-based synthetic data generation method, tailoring it for imbalanced regression. The new method integrates relevance and density-based mechanisms to guide sampling in sparse regions of the target space and employs a threshold-free, feature-driven generation process. Our experimental study focuses on the prediction of extreme target values across benchmark datasets. The results indicate that the proposed method is competitive with other resampling and generative strategies in terms of performance, while offering faster execution and greater transparency. These results highlight the method's potential as a transparent, scalable data-level strategy for improving regression models in imbalanced domains. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:42:20 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Pinheiro",
"António Pedro",
""
],
[
"Ribeiro",
"Rita P.",
""
]
] |
2506.02812 | Muhammad Zain Mobeen | Muhammad Zain Mobeen, Tomasz Kami\'nski, Stephen Potter | Constraining the physical structure of the circumstellar environment of V838 Monocerotis remnant | null | null | null | null | astro-ph.SR astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | V838 Monocerotis (V838 Mon) erupted in 2002 as a luminous red nova after which it cooled and began to form dust. The remnant is predicted to become a blue straggler. Interferometric observations in the HKLM bands have uncovered a stable bipolar feature in the closest vicinity of the remnant star. We aim to constrain the physical structure and nature of the circumstellar material immediately surrounding V838 Mon. Using the radiative-transfer code RADMC3D we managed to constrain the dust density distribution that best represents recent VLTI and CHARA interferometric imaging experiments. We also present recent SAAO-HIPPO polarimetric measurements which further test dust distribution models. We find that a multi-component model consisting of jets, a torus, and an ellipsoid provides an adequate fit to H band interferometric observables, however, it struggles to reproduce the extremely small closure phase deviations in the K band. The jets in our model are vital to produce nonzero closure phases, as they help to produce a gap in the torus, which is the sole source of asymmetry. The polarimetric measurements show that the intrinsic linear polarization is currently very low, with degree of polarization < 2%, consistent with our model. Although not unique, our dust model suggests a persistent torus-like structure and jet or jets in the remnant, in agreement with several predictions for an intermediate-mass merger product. However, even though these features may be important signposts of the evolution of V838 Mon to the blue straggler phase, the origin of these features, remains largely unclear. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:42:56 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Mobeen",
"Muhammad Zain",
""
],
[
"Kamiński",
"Tomasz",
""
],
[
"Potter",
"Stephen",
""
]
] |
2506.02813 | Jascha Achterberg Mr | Jack Cook, Danyal Akarca, Rui Ponte Costa, Jascha Achterberg | Brain-Like Processing Pathways Form in Models With Heterogeneous Experts | 25 pages, 14 figures | null | null | null | q-bio.NC cs.NE | http://creativecommons.org/licenses/by/4.0/ | The brain is made up of a vast set of heterogeneous regions that dynamically organize into pathways as a function of task demands. Examples of such pathways can be seen in the interactions between cortical and subcortical networks during learning. This raises the question of how exactly brain regions organize into these dynamic groups. In this work, we use an extension of the Heterogeneous Mixture-of-Experts architecture, to show that heterogeneous regions do not form processing pathways by themselves, implying that the brain likely implements specific constraints which result in reliable formation of pathways. We identify three biologically relevant inductive biases that encourage pathway formation: a routing cost imposed on the use of more complex regions, a scaling factor that reduces this cost when task performance is low, and randomized expert dropout. When comparing our resulting Mixture-of-Pathways model with the brain, we observe that the artificial pathways match how the brain uses cortical and subcortical systems to learn and solve tasks of varying difficulty. In summary, we introduce a novel framework for investigating how the brain forms task-specific pathways through inductive biases which may make Mixture-of-Experts architectures in general more adaptive. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:43:13 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Cook",
"Jack",
""
],
[
"Akarca",
"Danyal",
""
],
[
"Costa",
"Rui Ponte",
""
],
[
"Achterberg",
"Jascha",
""
]
] |
2506.02815 | Anne Poot | Anne Poot, Iuri Rocha, Pierre Kerfriden, Frans van der Meeer | The Bayesian Finite Element Method in Inverse Problems: a Critical Comparison between Probabilistic Models for Discretization Error | 20 pages, 9 figures | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | When using the finite element method (FEM) in inverse problems, its discretization error can produce parameter estimates that are inaccurate and overconfident. The Bayesian finite element method (BFEM) provides a probabilistic model for the epistemic uncertainty due to discretization error. In this work, we apply BFEM to various inverse problems, and compare its performance to the random mesh finite element method (RM-FEM) and the statistical finite element method (statFEM), which serve as a frequentist and inference-based counterpart to BFEM. We find that by propagating this uncertainty to the posterior, BFEM can produce more accurate parameter estimates and prevent overconfidence, compared to FEM. Because the BFEM covariance operator is designed to leave uncertainty only in the appropriate space, orthogonal to the FEM basis, BFEM is able to outperform RM-FEM, which does not have such a structure to its covariance. Although inferring the discretization error via a model misspecification component is possible as well, as is done in statFEM, the feasibility of such an approach is contingent on the availability of sufficient data. We find that the BFEM is the most robust way to consistently propagate uncertainty due to discretization error to the posterior of a Bayesian inverse problem. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:45:45 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Poot",
"Anne",
""
],
[
"Rocha",
"Iuri",
""
],
[
"Kerfriden",
"Pierre",
""
],
[
"van der Meeer",
"Frans",
""
]
] |
2506.02816 | Luca Bergamaschi Prof. | L. Bergamaschi and A. Martinez and J.W. Pearson and A. Potschka | Eigenvalue bounds for preconditioned symmetric multiple saddle-point matrices | null | null | null | null | math.NA cs.NA math.OC | http://creativecommons.org/licenses/by/4.0/ | We develop eigenvalue bounds for symmetric, block tridiagonal multiple saddle-point linear systems, preconditioned with block diagonal matrices. We extend known results for $3 \times 3$ block systems [Bradley and Greif, IMA J.\ Numer. Anal. 43 (2023)] and for $4 \times 4$ systems [Pearson and Potschka, IMA J. Numer. Anal. 44 (2024)] to an arbitrary number of blocks. Moreover, our results generalize the bounds in [Sogn and Zulehner, IMA J. Numer. Anal. 39 (2018)], developed for an arbitrary number of blocks with null diagonal blocks. Extension to the bounds when the Schur complements are approximated is also provided, using perturbation arguments. Practical bounds are also obtained for the double saddle-point linear system. Numerical experiments validate our findings. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:46:26 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Bergamaschi",
"L.",
""
],
[
"Martinez",
"A.",
""
],
[
"Pearson",
"J. W.",
""
],
[
"Potschka",
"A.",
""
]
] |
2506.02817 | Sabiju Valappil | Sabiju Valiya Valappil, Peter Harmsma, Maurits van der Heiden, Martin Verweij, Paul van Neer | Semi-analytical modeling of receive transfer function and thermal noise of integrated photonic ultrasound transducers | null | null | null | null | physics.app-ph physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ultrasound transducers (UTs) are extensively used in several applications across a multitude of disciplines. A new type of UTs namely integrated photonic ultrasound transducers (IPUTs) possess superior performance due to the presence of optical interrogation systems, avoiding electric crosstalk and thermal electronic noise of the sensor. However, a major component of the IPUT's noise floor is its thermal acoustic noise. Several studies have been proposed to characterize IPUTs' behavior; nevertheless, these are either incomplete (model only the thermal noise) or targeted to characterize specific responses such as static behavior, in which the modeled receive transfer function (RTF) is about two orders lower than the experiments. In this study, we develop semi-analytical models based on time-domain finite element analysis and analytical expressions to characterize the RTF and thermal noise-induced noise equivalent pressure of IPUTs. We validate the models by comparing them with the literature, where we obtain a close match between them. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:47:08 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Valappil",
"Sabiju Valiya",
""
],
[
"Harmsma",
"Peter",
""
],
[
"van der Heiden",
"Maurits",
""
],
[
"Verweij",
"Martin",
""
],
[
"van Neer",
"Paul",
""
]
] |
2506.02818 | Ekaterina Grishina | Ekaterina Grishina, Mikhail Gorbunov, Maxim Rakhuba | ProcrustesGPT: Compressing LLMs with Structured Matrices and Orthogonal Transformations | Accepted by ACL Findings | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) demonstrate impressive results in natural language processing tasks but require a significant amount of computational and memory resources. Structured matrix representations are a promising way for reducing the number of parameters of these models. However, it seems unrealistic to expect that weight matrices of pretrained models can be accurately represented by structured matrices without any fine-tuning. To overcome this issue, we utilize the fact that LLM output is invariant under certain orthogonal transformations of weight matrices. This insight can be leveraged to identify transformations that significantly improve the compressibility of weights within structured classes. The proposed approach is applicable to various types of structured matrices that support efficient projection operations. Code is available at https://github.com/GrishKate/ProcrustesGPT | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:47:23 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Grishina",
"Ekaterina",
""
],
[
"Gorbunov",
"Mikhail",
""
],
[
"Rakhuba",
"Maxim",
""
]
] |
2506.02819 | Ryo Kato | Ryo Kato and Keitaro Takahashi | Realistic assessment of a single gravitational wave source localization taking into account precise pulsar distances with pulsar timing arrays | 18 pages, 6 figures | null | null | null | gr-qc astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pulsar timing arrays (PTAs) are anticipated to detect continuous gravitational waves (GWs) from individual supermassive black hole binaries (SMBHBs) in the near future. To identify the host galaxy of a GW source, PTAs require significantly improved angular resolution beyond the typical range of 100-1000 square degrees achieved by recent continuous GW searches. In this study, we investigate how precise pulsar distance measurements can enhance the localization of a single GW source. Accurate distance information, comparable to or better than the GW wavelength (typically 1~pc) can refine GW source localization. In the near future, with the advent of Square Kilometre Array (SKA), such high-precision distance measurements will be feasible for a few nearby pulsars. We focus on the relatively nearby pulsars J0437-4715 (156 pc) and J0030+0451 (331 pc), incorporating their actual distance uncertainties based on current VLBI measurements and the anticipated precision of the SKA-era. By simulating 87 pulsars with the GW signal and Gaussian white noise in the timing residuals, we assess the impact of the pulsar distance information on GW source localization. Our results show that without precise distance information, localization remains insufficient to identify host galaxies under 10 ns noise. However, incorporating SKA-era distance precision for nearby pulsars J0437-4715 and J0030+0451 can reduce localization uncertainties to the required level of $10^{-3}$ $\rm deg^{2}$. Localization accuracy strongly depends on the geometric configuration of pulsars with well-measured distances and improves notably near and between such pulsars. The improvement of the localization will greatly aid in identifying the host galaxy of a GW source and constructing an SMBHB catalog. It will further enable follow-up electromagnetic observations to investigate the SMBHB in greater detail. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:48:08 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kato",
"Ryo",
""
],
[
"Takahashi",
"Keitaro",
""
]
] |
2506.02820 | Marine Aulnette | Marine Aulnette, Noa Burshtein, Arash Alizad Banaei, Luca Brandt, Simon J. Haward, Amy Q. Shen, Blaise Delmotte and Anke Lindner | Transport of spherical microparticles in a 3D vortex flow | 20 pages, 8 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | Particles are common in biological and environmental flows and are widely used in industrial and pharmaceutical applications. Their motion and flow dynamics are strongly affected by interactions with the surrounding flow structure. While particle-flow interactions have been extensively studied in low Reynolds number (Re) flows as well as in fully developed turbulence, the transport mechanisms of these particles in intermediate flow regimes remain less explored. Here, we investigate the response of neutrally buoyant spherical particles to a single vortex flow field. Using a microfluidic cross-slot geometry, we generate a well-characterized, stationary, three-dimensional streamwise vortex at moderate $\text{Re}$ ($\sim 50$). Our experimental results, supported by numerical simulations, show that with increasing particle diameter, they are progressively excluded from the vortex core. Initially, small particles follow a Burgers vortex-like self-similar motion, but for larger particle diameters, deviations from this trend emerge due to fluid inertia and finite-size effects. These findings enhance our understanding of particle dynamics in vortical flows and have implications for microfluidic applications involving particle sorting and separation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:48:48 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Aulnette",
"Marine",
""
],
[
"Burshtein",
"Noa",
""
],
[
"Banaei",
"Arash Alizad",
""
],
[
"Brandt",
"Luca",
""
],
[
"Haward",
"Simon J.",
""
],
[
"Shen",
"Amy Q.",
""
],
[
"Delmotte",
"Blaise",
""
],
[
"Lindner",
"Anke",
""
]
] |
2506.02821 | Sajad Ahanger | Sajad Ahanger, Sunder Sahayanathan, Sitha K. Jagan, Shah Zahir and Naseer Iqbal | Multi-wavelength analysis of FSRQ B2 1348+30B: Constraints on the jet power | 14 pages, 6 figures, 5 tables, Accepted for publication in JHEAP | null | null | null | astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present 14.5-year multi-wavelength analysis of flat-spectrum radio quasar B2 1348+30B using Swift-UVOT, Swift-XRT, and Fermi-LAT observations. In the gamma-ray band, the 3 day bin lightcurve reveals two major flaring events on 2010-09-19 (55458 MJD) and 2022-05-26 (59725 MJD) detected at flux levels $(2.5\pm 0.5) \times 10^{-7}\,\rm{ph\,cm^{-2}\,s^{-1}}$ and $(5.2\pm 0.6) \times 10^{-7}\,\rm{ph\,cm^{-2}\,s^{-1}}$. The Bayesian block analysis of the flares suggested the variability timescale to be $\leq$ 3\,day. To study the dynamic nature of the source, multi-wavelength spectrum was obtained for three flux states which includes the two flaring state and a relative low state. The $\gamma$-ray spectra of the source in all the states are well fitted by a power-law model with maximum photon energy < 20 GeV. In X-ray, a power-law model can explain the flaring state spectra while a broken power-law with extremely hard high energy component was required to model the low flux state. This indicates the presence of the low energy cutoff in the Compton spectral component. A simple one-zone leptonic model involving synchrotron, synchrotron self Compton and external Compton mechanism can successfully reproduce the broad-band spectral energy distribution of all the flux states. The model parameters suggest significant increase in the jet Lorentz factor during the high flux states. Further, the best fit parameters are used to constrain the minimum energy of the emitting electron distribution from the hard high energy spectrum of the low flux state. This analysis was extended to draw limits on the kinetic power of the blazar jet and was compared with the Eddington luminosity of the central black hole. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:50:06 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Ahanger",
"Sajad",
""
],
[
"Sahayanathan",
"Sunder",
""
],
[
"Jagan",
"Sitha K.",
""
],
[
"Zahir",
"Shah",
""
],
[
"Iqbal",
"Naseer",
""
]
] |
2506.02822 | Sanjib Dey | Duttatreya, Sanjib Dey | Enhanced quantum phase estimation with $q$-deformed nonideal nonclassical light | 8 pages, 4 figures | null | null | null | quant-ph hep-th math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | We investigate quantum phase estimation in a Mach-Zehnder interferometer using q-deformed photon states, including q-coherent and q-cat states, which model realistic deviations from ideal light sources. By deriving closed-form photon count likelihoods via the Jordan-Schwinger mapping, we compute the quantum and classical Fisher information and perform Bayesian inference on simulated detector data. Our results show that photon counting remains an optimal measurement strategy even for deformed states, with classical and quantum Fisher information in exact agreement. Furthermore, the phase sensitivity improves with increasing q-deformation, indicating enhanced metrological performance driven by nonclassical photon statistics. These findings highlight the utility of q-deformed states in quantum sensing. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:50:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Duttatreya",
"",
""
],
[
"Dey",
"Sanjib",
""
]
] |
2506.02823 | Jingdie Xin | Jingdie Xin, Yan Wang, Feng Shu, Feng Zhao, Yifan Zhao, Hao Jiang | Energy Efficiency Analysis of Active RIS-enhanced Wireless Network under Power-Sum Constraint | null | null | null | null | cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, as a green wireless technology, active reconfigurable intelligent surface (RIS) attracts numerous research activities due to its amplifying ability to combat the double-fading effect compared to passive one. How about its energy efficiency (EE) over passive one? Below, the EE of active RIS-aided wireless network in Rayleigh fading channels is analyzed. Using the law of large numbers, EE is derived as a function of five factors: power allocation factor, the number (N) of RIS elements, the total power, the noise variances at RIS and at user. To evaluate each factor's impact, the simple EE function for the concerning factor is given with others fixed. To assess the impact of N on EE, we establish an equation with the EE of active RIS equaling that of passive one, and three methods, bisection, Newton's method, and simulated annealing, are designed to find the roots of this equation. Simulation results show that as N tends to medium-scale or large-scale, the asymptotic performance formula is consistent with the exact EE expression well. As N varies from small-scale to large-scale, the active RIS intersects passive one at some point. When N< N_0, active RIS performs better than passive one in terms of EE. Otherwise, there is a converse conclusion. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:51:09 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Xin",
"Jingdie",
""
],
[
"Wang",
"Yan",
""
],
[
"Shu",
"Feng",
""
],
[
"Zhao",
"Feng",
""
],
[
"Zhao",
"Yifan",
""
],
[
"Jiang",
"Hao",
""
]
] |
2506.02824 | Huazhi Dong | Huazhi Dong, Ronald B. Liu, Sihao Teng, Delin Hu, Peisan (Sharel) E, Francesco Giorgio-Serchi, and Yunjie Yang | Efficient Tactile Perception with Soft Electrical Impedance Tomography and Pre-trained Transformer | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tactile sensing is fundamental to robotic systems, enabling interactions through physical contact in multiple tasks. Despite its importance, achieving high-resolution, large-area tactile sensing remains challenging. Electrical Impedance Tomography (EIT) has emerged as a promising approach for large-area, distributed tactile sensing with minimal electrode requirements which can lend itself to addressing complex contact problems in robotics. However, existing EIT-based tactile reconstruction methods often suffer from high computational costs or depend on extensive annotated simulation datasets, hindering its viability in real-world settings. To address this shortcoming, here we propose a Pre-trained Transformer for EIT-based Tactile Reconstruction (PTET), a learning-based framework that bridges the simulation-to-reality gap by leveraging self-supervised pretraining on simulation data and fine-tuning with limited real-world data. In simulations, PTET requires 99.44 percent fewer annotated samples than equivalent state-of-the-art approaches (2,500 vs. 450,000 samples) while achieving reconstruction performance improvements of up to 43.57 percent under identical data conditions. Fine-tuning with real-world data further enables PTET to overcome discrepancies between simulated and experimental datasets, achieving superior reconstruction and detail recovery in practical scenarios. The improved reconstruction accuracy, data efficiency, and robustness in real-world tasks establish it as a scalable and practical solution for tactile sensing systems in robotics, especially for object handling and adaptive grasping under varying pressure conditions. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:53:13 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Dong",
"Huazhi",
"",
"Sharel"
],
[
"Liu",
"Ronald B.",
"",
"Sharel"
],
[
"Teng",
"Sihao",
"",
"Sharel"
],
[
"Hu",
"Delin",
"",
"Sharel"
],
[
"Peisan",
"",
"",
"Sharel"
],
[
"E",
"",
""
],
[
"Giorgio-Serchi",
"Francesco",
""
],
[
"Yang",
"Yunjie",
""
]
] |
2506.02825 | Vince Lyzinski | Tong Qi, Vera Andersson, Peter Viechnicki, Vince Lyzinski | Asymptotically perfect seeded graph matching without edge correlation (and applications to inference) | 10 figures, 35 pages | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the OmniMatch algorithm for seeded multiple graph matching. In the setting of $d$-dimensional Random Dot Product Graphs (RDPG), we prove that under mild assumptions, OmniMatch with $s$ seeds asymptotically and efficiently perfectly aligns $O(s^{\alpha})$ unseeded vertices -- for $\alpha<2\wedge d/4$ -- across multiple networks even in the presence of no edge correlation. We demonstrate the effectiveness of our algorithm across numerous simulations and in the context of shuffled graph hypothesis testing. In the shuffled testing setting, testing power is lost due to the misalignment/shuffling of vertices across graphs, and we demonstrate the capacity of OmniMatch to correct for misaligned vertices prior to testing and hence recover the lost testing power. We further demonstrate the algorithm on a pair of data examples from connectomics and machine translation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:54:24 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Qi",
"Tong",
""
],
[
"Andersson",
"Vera",
""
],
[
"Viechnicki",
"Peter",
""
],
[
"Lyzinski",
"Vince",
""
]
] |
2506.02826 | Kevin Ollmann | Kevin Ollmann, Florian Kirchschlager, Thomas A. Stuber, Katsiaryna Tsishchankava, Alexis Matter, Steve Ertel, Tim D. Pearce, Alexander V. Krivov, Sebastian Wolf | Hot exozodiacal dust around Fomalhaut: The MATISSE perspective | Accepted for a publication in Astronomy & Astrophysics | null | 10.1051/0004-6361/202451592 | null | astro-ph.EP astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Excess over the stellar photospheric emission of main-sequence stars has been found in interferometric near-infrared observations, and is attributed to the presence of hot exozodiacal dust (HEZD). As part of our effort to detect and characterize HEZD around the nearby A3 V star Fomalhaut, we carried out the first interferometric observations with the MATISSE instrument at the VLTI in the photometric bands L and M for the Fomalhaut system. Assuming a dust distribution either as a narrow ring or spherical shell for modeling the HEZD, we aim to constrain the HEZD parameters by generating visibilities and fitting them to the MATISSE data using different approaches. The L band data provide a marginal detection of circumstellar radiation, potentially caused by the presence of HEZD, which is only the second detection of HEZD emission in the L band. An analysis of the data with different fitting approaches showed that the best-fit values for the HEZD parameters are consistent with those of previous Fomalhaut observations, which underlines the functionality of MATISSE. Assuming a dust ring, it would have an inner ring radius of \(0.11\ \mathrm{au}\), an outer ring radius of \(0.12\ \mathrm{au}\), a narrow dust grain size distribution around a dust grain radius constrained by \(0.53\ \mu\mathrm{m}\), and a total dust mass of \(3.25\times 10^{-10}\ \rm{M}_{\oplus}\). Finally, the results indicate that the choice of the geometric model has a more significant impact on the derived dust-to-star flux ratio than the specific fitting approach applied. Since different dust-to-star flux ratios can result from the applied fitting approaches, this also has an impact on the parameter values of the HEZD around Fomalhaut and most likely for other HEZD systems. Moreover, further NIR and MIR data are required for a more comprehensive description of the emission originating in the vicinity of Fomalhaut. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:55:50 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Ollmann",
"Kevin",
""
],
[
"Kirchschlager",
"Florian",
""
],
[
"Stuber",
"Thomas A.",
""
],
[
"Tsishchankava",
"Katsiaryna",
""
],
[
"Matter",
"Alexis",
""
],
[
"Ertel",
"Steve",
""
],
[
"Pearce",
"Tim D.",
""
],
[
"Krivov",
"Alexander V.",
""
],
[
"Wolf",
"Sebastian",
""
]
] |
2506.02827 | Yulin Dou | Yulin Dou, Jiangming Liu | TO-GATE: Clarifying Questions and Summarizing Responses with Trajectory Optimization for Eliciting Human Preference | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) can effectively elicit human preferences through multi-turn dialogue. Complex tasks can be accomplished through iterative clarifying questions and final responses generated by an LLM acting as a questioner (STaR-GATE; Andukuri et al., 2024}). However, existing approaches based on self-taught reasoning struggle to identify optimal dialogue trajectories and avoid irrelevant questions to the tasks. To address this limitation, we propose TO-GATE, a novel framework that enhances question generation through trajectory optimization, which consists of two key components: a clarification resolver that generates optimal questioning trajectories, and a summarizer that ensures task-aligned final responses. The trajectory optimization enables the model to produce effective elicitation questions and summary responses tailored to specific tasks. Experimental results demonstrate that TO-GATE significantly outperforms baseline methods, achieving a 9.32% improvement on standard preference elicitation tasks. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:58:07 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Dou",
"Yulin",
""
],
[
"Liu",
"Jiangming",
""
]
] |
2506.02828 | Ahmet Burak Ozyurt | Ahmet Burak Ozyurt and John S. Thompson | Target Sensing Performance in Disaster-Specific ISAC Networks | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | As sixth-generation (6G) wireless technology emerges, integrated sensing and communication (ISAC) networks offer significant potential for enhancing real-time monitoring in disaster areas. However, existing ISAC approaches often fail to address the unique challenges of dynamic and cluttered disaster areas, resulting in limited sensing coverage and interruptions in sensing service. To address these limitations, this work proposes a mobile ISAC network specifically designed for disaster scenarios. By leveraging stochastic geometry, we derive closed-form expressions for sensing coverage and introduce a novel performance metric to evaluate sensing service continuity. Simulation results validate the analytical derivations and offer key insights into network design. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:58:31 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Ozyurt",
"Ahmet Burak",
""
],
[
"Thompson",
"John S.",
""
]
] |
2506.02829 | Daniel Loughran | D. R. Heath-Brown and Daniel Loughran | Manin's conjecture for quintic del Pezzo surfaces with a conic bundle structure | 98 pages | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate Manin's conjecture for del Pezzo surfaces of degree five with a conic bundle structure, proving matching upper and lower bounds, and the full conjecture in the Galois general case. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:58:57 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Heath-Brown",
"D. R.",
""
],
[
"Loughran",
"Daniel",
""
]
] |
2506.02830 | Maximilian Weisenseel | Maximilian Weisenseel, Julia Andersen, Samira Akili, Christian Imenkamp, Hendrik Reiter, Christoffer Rubensson, Wilhelm Hasselbring, Olaf Landsiedel, Xixi Lu, Jan Mendling, Florian Tschorsch, Matthias Weidlich, Agnes Koschmider | Process Mining on Distributed Data Sources | null | null | null | null | cs.ET cs.DB cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Major domains such as logistics, healthcare, and smart cities increasingly rely on sensor technologies and distributed infrastructures to monitor complex processes in real time. These developments are transforming the data landscape from discrete, structured records stored in centralized systems to continuous, fine-grained, and heterogeneous event streams collected across distributed environments. As a result, traditional process mining techniques, which assume centralized event logs from enterprise systems, are no longer sufficient. In this paper, we discuss the conceptual and methodological foundations for this emerging field. We identify three key shifts: from offline to online analysis, from centralized to distributed computing, and from event logs to sensor data. These shifts challenge traditional assumptions about process data and call for new approaches that integrate infrastructure, data, and user perspectives. To this end, we define a research agenda that addresses six interconnected fields, each spanning multiple system dimensions. We advocate a principled methodology grounded in algorithm engineering, combining formal modeling with empirical evaluation. This approach enables the development of scalable, privacy-aware, and user-centric process mining techniques suitable for distributed environments. Our synthesis provides a roadmap for advancing process mining beyond its classical setting, toward a more responsive and decentralized paradigm of process intelligence. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 12:59:36 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Weisenseel",
"Maximilian",
""
],
[
"Andersen",
"Julia",
""
],
[
"Akili",
"Samira",
""
],
[
"Imenkamp",
"Christian",
""
],
[
"Reiter",
"Hendrik",
""
],
[
"Rubensson",
"Christoffer",
""
],
[
"Hasselbring",
"Wilhelm",
""
],
[
"Landsiedel",
"Olaf",
""
],
[
"Lu",
"Xixi",
""
],
[
"Mendling",
"Jan",
""
],
[
"Tschorsch",
"Florian",
""
],
[
"Weidlich",
"Matthias",
""
],
[
"Koschmider",
"Agnes",
""
]
] |
2506.02831 | Dylan Grigg | Dylan Grigg, Steven Tingay and Marcin Sokolowski | The Growing Impact of Unintended Starlink Broadband Emission on Radio Astronomy in the SKA-Low Frequency Range | 8 pages, 5 figures, 1 table | null | null | null | astro-ph.IM astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | We present the largest survey to date characterising intended and unintended emission from Starlink satellites across the SKA-Low frequency range. This survey analyses ~76 million full sky images captured over ~29 days of observing with an SKA-Low prototype station - the Engineering Development Array 2 - at the site of SKA-Low. We report 112,534 individual detections of 1,806 unique Starlink satellites, some emitting broadband emission and others narrowband emission. Our analysis compares observations across different models of Starlink satellite, with 76% of all v2-mini Ku and 71% of all v2-mini Direct to Cell satellites identified. It is shown that in the worst cases, some datasets have a detectable Starlink satellite in ~30% of all images acquired. Emission from Starlink satellites is detected in primary and secondary frequency ranges protected by the International Telecommunication Union, with 13 satellites identified between 73.00 - 74.60 MHz and 703 identified between 150.05 - 153.00 MHz. We also detect the reflections of terrestrial FM radio off different models of Starlink satellites at 99.70 MHz. The polarisation of the broadband emission shows the flux density of two orthogonal polarisations is anti-correlated with temporally shifting spectral structure observed. We compare our results to previous EDA2 and LOFAR results and provide open public access to our final data products to assist in quantifying future changes in this emission. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:00:11 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Grigg",
"Dylan",
""
],
[
"Tingay",
"Steven",
""
],
[
"Sokolowski",
"Marcin",
""
]
] |
2506.02832 | Arseny Mironov | A. A. Mironov, S. S. Bulanov, A. Di Piazza, M. Grech, L. Lancia, S. Meuren, J. Palastro, C. Riconda, H. G. Rinderknecht, P. Tzeferacos, and G. Gregori | Testing strong-field QED with the avalanche precursor | 17 pages, 8 figures, 3 tables | null | null | null | physics.plasm-ph hep-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A two-beam high-power laser facility is essential for the study of one of the most captivating phenomena predicted by strong-field quantum electrodynamics (QED) and yet unobserved experimentally: the avalanche-type cascade. In such a cascade, the energy of intense laser light can be efficiently transformed into high-energy radiation and electron-positron pairs. The future 50-petawatt-scale laser facility NSF OPAL will provide unique opportunities for studying such strong-field QED effects, as it is designed to deliver two ultra-intense, tightly focused laser pulses onto the interaction point. In this work, we investigate the potential of such a facility for studying elementary particle and plasma dynamics deeply in the quantum radiation-dominated regime, and the generation of QED avalanches. With 3D particle-in-cell simulations, we demonstrate that QED avalanche precursors can be reliably triggered under realistic laser parameters and layout (namely, focusing $f/2$, tilted optical axes, and non-ideal co-pointing) with the anticipated capabilities of NSF OPAL. We demonstrate that seed electrons can be efficiently injected into the laser focus by using targets of three types: a gas of heavy atoms, an overcritical plasma, and a thin foil. A strong positron and high-energy photon signal is generated in all cases. The cascade properties can be identified from the final particle distributions, which have a clear directional pattern. At increasing laser field intensity, such distributions provide signatures of the transition, first, to the radiation-dominated interaction regime, and then to a QED avalanche. Our findings can also be used for designing related future experiments. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:00:14 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Mironov",
"A. A.",
""
],
[
"Bulanov",
"S. S.",
""
],
[
"Di Piazza",
"A.",
""
],
[
"Grech",
"M.",
""
],
[
"Lancia",
"L.",
""
],
[
"Meuren",
"S.",
""
],
[
"Palastro",
"J.",
""
],
[
"Riconda",
"C.",
""
],
[
"Rinderknecht",
"H. G.",
""
],
[
"Tzeferacos",
"P.",
""
],
[
"Gregori",
"G.",
""
]
] |
2506.02833 | Aki Pulkkinen | Aki Pulkkinen, Veenavee Nipunika Kothalawala, Kosuke Suzuki, Bernardo Barbiellini, Johannes Nokelainen, Wei-Chi Chiu, Bahadur Singh, Hsin Lin, Alok K. Pandey, Naoaki Yabuuchi, Naruki Tsuji, Yoshiharu Sakurai, Hiroshi Sakurai, J\'an Min\'ar, Arun Bansil | Probing the semiconductor-to-dirac semimetal transition in Na-Sb-Bi alloys with x-ray Compton scattering | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | We discuss electron redistribution during the Semiconductor-to-Dirac Semimetal transition in Na-Sb-Bi alloys using x-ray Compton scattering experiments combined with first-principles electronic structure modeling. A robust signature of the semiconductor-to-Dirac semimetal transition is identified in the spherically averaged Compton profile. We demonstrate how the number of electrons involved in this transition can be estimated to provide a novel descriptor for quantifying the strength of spin-orbit coupling responsible for driving the transition. The associated theoretical deviation of the Born charge of Na in Na$_3$Bi from the expected ionic charge of +1 is found to be consistent with the corresponding experimental value of about 10%. Our study also shows the sensitivity of the Compton scattering technique toward capturing the spillover of Bi 6p relativistic states onto Na sites. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:01:43 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Pulkkinen",
"Aki",
""
],
[
"Kothalawala",
"Veenavee Nipunika",
""
],
[
"Suzuki",
"Kosuke",
""
],
[
"Barbiellini",
"Bernardo",
""
],
[
"Nokelainen",
"Johannes",
""
],
[
"Chiu",
"Wei-Chi",
""
],
[
"Singh",
"Bahadur",
""
],
[
"Lin",
"Hsin",
""
],
[
"Pandey",
"Alok K.",
""
],
[
"Yabuuchi",
"Naoaki",
""
],
[
"Tsuji",
"Naruki",
""
],
[
"Sakurai",
"Yoshiharu",
""
],
[
"Sakurai",
"Hiroshi",
""
],
[
"Minár",
"Ján",
""
],
[
"Bansil",
"Arun",
""
]
] |
2506.02835 | Hyeongjun Kim | Hyeongjun Kim, Hyunsik Oh, Jeongsoo Park, Yunho Kim, Donghoon Youm, Moonkyu Jung, Minho Lee, Jemin Hwangbo | High-speed control and navigation for quadrupedal robots on complex and discrete terrain | null | Science Robotics 10.102 (2025): eads6192 | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | High-speed legged navigation in discrete and geometrically complex environments is a challenging task because of the high-degree-of-freedom dynamics and long-horizon, nonconvex nature of the optimization problem. In this work, we propose a hierarchical navigation pipeline for legged robots that can traverse such environments at high speed. The proposed pipeline consists of a planner and tracker module. The planner module finds physically feasible foothold plans by sampling-based optimization with fast sequential filtering using heuristics and a neural network. Subsequently, rollouts are performed in a physics simulation to identify the best foothold plan regarding the engineered cost function and to confirm its physical consistency. This hierarchical planning module is computationally efficient and physically accurate at the same time. The tracker aims to accurately step on the target footholds from the planning module. During the training stage, the foothold target distribution is given by a generative model that is trained competitively with the tracker. This process ensures that the tracker is trained in an environment with the desired difficulty. The resulting tracker can overcome terrains that are more difficult than what the previous methods could manage. We demonstrated our approach using Raibo, our in-house dynamic quadruped robot. The results were dynamic and agile motions: Raibo is capable of running on vertical walls, jumping a 1.3-meter gap, running over stepping stones at 4 meters per second, and autonomously navigating on terrains full of 30{\deg} ramps, stairs, and boxes of various sizes. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:04:26 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kim",
"Hyeongjun",
""
],
[
"Oh",
"Hyunsik",
""
],
[
"Park",
"Jeongsoo",
""
],
[
"Kim",
"Yunho",
""
],
[
"Youm",
"Donghoon",
""
],
[
"Jung",
"Moonkyu",
""
],
[
"Lee",
"Minho",
""
],
[
"Hwangbo",
"Jemin",
""
]
] |
2506.02836 | Jan O. Bauer | Maria Laura Battagliola and Jan O. Bauer | Localized Functional Principal Component Analysis Based on Covariance Structure | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Functional principal component analysis (FPCA) is a widely used technique in functional data analysis for identifying the primary sources of variation in a sample of random curves. The eigenfunctions obtained from standard FPCA typically have non-zero support across the entire domain. In applications, however, it is often desirable to analyze eigenfunctions that are non-zero only on specific portions of the original domain-and exhibit zero regions when little is contributed to a specific direction of variability-allowing for easier interpretability. Our method identifies sparse characteristics of the underlying stochastic process and derives localized eigenfunctions by mirroring these characteristics without explicitly enforcing sparsity. Specifically, we decompose the stochastic process into uncorrelated sub-processes, each supported on disjoint intervals. Applying FPCA to these sub-processes yields localized eigenfunctions that are naturally orthogonal. In contrast, approaches that enforce localization through penalization must additionally impose orthogonality. Moreover, these approaches can suffer from over-regularization, resulting in eigenfunctions and eigenvalues that deviate from the inherent structure of their population counterparts, potentially misrepresenting data characteristics. Our approach avoids these issues by preserving the inherent structure of the data. Moreover, since the sub-processes have disjoint supports, the eigenvalues associated to the localized eigenfunctions allow for assessing the importance of each sub-processes in terms of its contribution to the total explained variance. We illustrate the effectiveness of our method through simulations and real data applications. Supplementary material for this article is available online. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:04:33 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Battagliola",
"Maria Laura",
""
],
[
"Bauer",
"Jan O.",
""
]
] |
2506.02837 | Zeenat Hameed | Zeenat Hameed, Chresten Traeholt | Optimal BESS Scheduling for Multi-Market Participation in the Nordics | Submitted to conference | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | Battery energy storage systems BESSs can provide fast frequency reserves and energy arbitrage in Nordic electricity markets but their limited energy capacity requires accurate revenue forecasts and coordinated bidding across multiple submarkets. This paper introduces a unified framework that employs generalized additive models GAMs to generate one week ahead forecasts of spot and FCRN revenues in Denmark, Finland, and Norway using three years of hourly price and volume data. Forecast outputs feed a stochastic mixed integer optimizer that co-optimizes BESS participation in FCRN, FCRD, spot markets, subject to state of charge constraints, inverter losses, and differing pay as bid and pay as clear rules. Comparative analyses evaluate forecast accuracy and quantify the impact of forecast errors on BESS bid acceptance, market selection, and profitability under realistic seasonal price patterns. Results demonstrate that spot markets exhibit consistently higher predictability than frequency markets, and that forecast errors modestly affect bid acceptance but do not alter overall market participation. The proposed approach provides BESS operators and investors with a tool to assess revenue uncertainty and optimize multi market strategies in Nordic power systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:05:39 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Hameed",
"Zeenat",
""
],
[
"Traeholt",
"Chresten",
""
]
] |
2506.02838 | Jizhou Wang | Jizhou Wang, Xiaodan Fang, Lei Huang, Yongfeng Huang | TaxAgent: How Large Language Model Designs Fiscal Policy | Accepted as oral presentation at ICME 2025 | null | null | null | cs.AI econ.GN q-fin.EC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Economic inequality is a global challenge, intensifying disparities in education, healthcare, and social stability. Traditional systems like the U.S. federal income tax reduce inequality but lack adaptability. Although models like the Saez Optimal Taxation adjust dynamically, they fail to address taxpayer heterogeneity and irrational behavior. This study introduces TaxAgent, a novel integration of large language models (LLMs) with agent-based modeling (ABM) to design adaptive tax policies. In our macroeconomic simulation, heterogeneous H-Agents (households) simulate real-world taxpayer behaviors while the TaxAgent (government) utilizes LLMs to iteratively optimize tax rates, balancing equity and productivity. Benchmarked against Saez Optimal Taxation, U.S. federal income taxes, and free markets, TaxAgent achieves superior equity-efficiency trade-offs. This research offers a novel taxation solution and a scalable, data-driven framework for fiscal policy evaluation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:06:19 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wang",
"Jizhou",
""
],
[
"Fang",
"Xiaodan",
""
],
[
"Huang",
"Lei",
""
],
[
"Huang",
"Yongfeng",
""
]
] |
2506.02839 | Yougang Lyu | Yougang Lyu, Xiaoyu Zhang, Lingyong Yan, Maarten de Rijke, Zhaochun Ren, Xiuying Chen | DeepShop: A Benchmark for Deep Research Shopping Agents | null | null | null | null | cs.IR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Web agents for online shopping have shown great promise in automating user interactions across e-commerce platforms. Benchmarks for assessing such agents do not reflect the complexity of real-world shopping scenarios, as they often consist of overly simple queries with deterministic paths, such as "Find iPhone 15." Real shopping scenarios are inherently more layered, involving multi-dimensional product attributes, search filters, and user-specific sorting preferences. To address this gap, we introduce DeepShop, a benchmark designed to evaluate web agents in complex and realistic online shopping environments. DeepShop comprises three key components. (1) Query diversity evolution: Starting from real user queries, we generate diverse queries across five popular online shopping domains. (2) Query complexity evolution: We further evolve these queries to increase complexity, considering product attributes, search filters, and sorting preferences, and classify them into three levels: easy, medium, and hard, based on the number of evolutions. (3) Fine-grained and holistic evaluation: We propose an automated evaluation framework that assesses agent performance in terms of fine-grained aspects (product attributes, search filters, and sorting preferences) and reports the overall success rate through holistic evaluation. We conduct a systematic evaluation of retrieval-augmented generation (RAG) methods, web agents, and deep research systems. Results show that RAG struggles with complex queries due to its lack of web interaction, while other methods face significant challenges with filters and sorting preferences, leading to low overall success rates. We also perform cross-category, complexity-based evaluations and error analyses to support the advancement of deep research shopping agents. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:08:17 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Lyu",
"Yougang",
""
],
[
"Zhang",
"Xiaoyu",
""
],
[
"Yan",
"Lingyong",
""
],
[
"de Rijke",
"Maarten",
""
],
[
"Ren",
"Zhaochun",
""
],
[
"Chen",
"Xiuying",
""
]
] |
2506.02840 | David Umsonst | David Umsonst and Mina Ferizbegovic | On dual-rate consensus under transmission delays | To be presented at the 2025 European Control Conference, 9 pages, 3 figures | null | null | null | eess.SY cs.SY math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate the problem of dual-rate consensus under transmission delays, where the control updates happen at a faster rate than the measurements being received. We assume that the measurements are delayed by a fixed delay and show that for all delays and rates, the system reaches a consensus if and only if the communication graph of the agents is connected and the control gain is chosen in a specific interval. Based on these results we dive deeper into the convergence properties and investigate how the convergence changes when we change the rate for sending measurements. We observe that in certain cases there exists a sweet spot for choosing the sampling rate of the measurements, which can improve the convergence to the consensus point. We then formulate an optimization problem to find a sampling rate to improve the convergence speed and provide a necessary and sufficient condition for the existence of a finite optimizer of this problem. Our results are verified with numerical simulations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:09:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Umsonst",
"David",
""
],
[
"Ferizbegovic",
"Mina",
""
]
] |
2506.02841 | Tom Danino | Tom Danino, Nahum Shimkin | Ensemble-MIX: Enhancing Sample Efficiency in Multi-Agent RL Using Ensemble Methods | null | null | null | null | eess.SY cs.LG cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-agent reinforcement learning (MARL) methods have achieved state-of-the-art results on a range of multi-agent tasks. Yet, MARL algorithms typically require significantly more environment interactions than their single-agent counterparts to converge, a problem exacerbated by the difficulty in exploring over a large joint action space and the high variance intrinsic to MARL environments. To tackle these issues, we propose a novel algorithm that combines a decomposed centralized critic with decentralized ensemble learning, incorporating several key contributions. The main component in our scheme is a selective exploration method that leverages ensemble kurtosis. We extend the global decomposed critic with a diversity-regularized ensemble of individual critics and utilize its excess kurtosis to guide exploration toward high-uncertainty states and actions. To improve sample efficiency, we train the centralized critic with a novel truncated variation of the TD($\lambda$) algorithm, enabling efficient off-policy learning with reduced variance. On the actor side, our suggested algorithm adapts the mixed samples approach to MARL, mixing on-policy and off-policy loss functions for training the actors. This approach balances between stability and efficiency and outperforms purely off-policy learning. The evaluation shows our method outperforms state-of-the-art baselines on standard MARL benchmarks, including a variety of SMAC II maps. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:13:15 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Danino",
"Tom",
""
],
[
"Shimkin",
"Nahum",
""
]
] |
2506.02842 | Stefano Fiorini | Stefano Fiorini, Hakan Aktas, Iulia Duta, Stefano Coniglio, Pietro Morerio, Alessio Del Bue, Pietro Li\`o | Sheaves Reloaded: A Directional Awakening | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sheaf Neural Networks (SNNs) represent a powerful generalization of Graph Neural Networks (GNNs) that significantly improve our ability to model complex relational data. While directionality has been shown to substantially boost performance in graph learning tasks and is key to many real-world applications, existing SNNs fall short in representing it. To address this limitation, we introduce the Directed Cellular Sheaf, a special type of cellular sheaf designed to explicitly account for edge orientation. Building on this structure, we define a new sheaf Laplacian, the Directed Sheaf Laplacian, which captures both the graph's topology and its directional information. This operator serves as the backbone of the Directed Sheaf Neural Network (DSNN), the first SNN model to embed a directional bias into its architecture. Extensive experiments on nine real-world benchmarks show that DSNN consistently outperforms baseline methods. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:13:56 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Fiorini",
"Stefano",
""
],
[
"Aktas",
"Hakan",
""
],
[
"Duta",
"Iulia",
""
],
[
"Coniglio",
"Stefano",
""
],
[
"Morerio",
"Pietro",
""
],
[
"Del Bue",
"Alessio",
""
],
[
"Liò",
"Pietro",
""
]
] |
2506.02843 | Shuai Yi | Shuai Yi, Yixiong Zou, Yuhua Li, Ruixuan Li | Random Registers for Cross-Domain Few-Shot Learning | Accepted by ICML 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cross-domain few-shot learning (CDFSL) aims to transfer knowledge from a data-sufficient source domain to data-scarce target domains. Although Vision Transformer (ViT) has shown superior capability in many vision tasks, its transferability against huge domain gaps in CDFSL is still under-explored. In this paper, we find an intriguing phenomenon: during the source-domain training, prompt tuning, as a common way to train ViT, could be harmful for the generalization of ViT in target domains, but setting them to random noises (i.e., random registers) could consistently improve target-domain performance. We then delve into this phenomenon for an interpretation. We find that learnable prompts capture domain information during the training on the source dataset, which views irrelevant visual patterns as vital cues for recognition. This can be viewed as a kind of overfitting and increases the sharpness of the loss landscapes. In contrast, random registers are essentially a novel way of perturbing attention for the sharpness-aware minimization, which helps the model find a flattened minimum in loss landscapes, increasing the transferability. Based on this phenomenon and interpretation, we further propose a simple but effective approach for CDFSL to enhance the perturbation on attention maps by adding random registers on the semantic regions of image tokens, improving the effectiveness and efficiency of random registers. Extensive experiments on four benchmarks validate our rationale and state-of-the-art performance. Codes and models are available at https://github.com/shuaiyi308/REAP. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:13:58 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Yi",
"Shuai",
""
],
[
"Zou",
"Yixiong",
""
],
[
"Li",
"Yuhua",
""
],
[
"Li",
"Ruixuan",
""
]
] |
2506.02844 | Ahmad Jafar Arifi | Ahmad Jafar Arifi, Ho-Meoyng Choi, Chueng-Ryong Ji | Beyond leading twist: $\rho$ meson decay constants and distribution amplitudes in a self-consistent light-front quark model | 14 pages, 2 figures | null | null | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | In this study, we present a comprehensive analysis of decay constants and chiral-even and chiral-odd distribution amplitudes (DAs) up to twist 4 for the $\rho$ meson in the standard light-front quark model (LFQM) based on the Bakamjian-Thomas (BT) construction. For the $\rho$ meson, which possesses both longitudinal $(h=0)$ and transverse $(h=\pm 1)$ polarizations, two types of decay constants, $f_{\rho}^{\parallel}$ and $f_{\rho}^{\perp}$, arises accordingly. We demonstrate that these decay constants can be self-consistently extracted from both local ($z^\mu=0$) and nonlocal ($z^\mu\neq 0$) matrix elements $\langle 0 | \bar{q}(z)\, \Gamma\, q(-z) | \rho(P,h) \rangle$, with $\Gamma = (\gamma^\mu, \sigma^{\mu\nu}, \gamma^\mu\gamma_5, \mathbf{1})$, in a manner independent of current components, polarizations, and reference frames. In particular, we emphasize the role of nonlocal matrix elements involving axial-vector and scalar currents, where mixing between $f_\rho^{\parallel}$ and $f_\rho^{\perp}$ occurs. We show that this mixing is consistently resolved through the BT construction, ensuring the proper extraction of these decay constants. Additionally, we investigate the structure of chiral-even DAs ($\phi_{2;\mathrm{V}}^\parallel, \phi_{3;\mathrm{V}}^\parallel, \psi_{3;\mathrm{A}}^\perp, \phi_{4;\mathrm{V}}^\parallel$) and chiral-odd DAs ($\phi_{2;\mathrm{T}}^\perp, \phi_{3;\mathrm{T}}^\perp, \psi_{3;\mathrm{S}}^\parallel, \phi_{4;\mathrm{T}}^\perp$) beyond leading twists, and present their corresponding $\xi$-moments and Gegenbauer moments. These results provide deeper insight into the nonperturbative structure of vector mesons and demonstrate the robustness and self-consistency of the LFQM based on the BT framework. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:14:54 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Arifi",
"Ahmad Jafar",
""
],
[
"Choi",
"Ho-Meoyng",
""
],
[
"Ji",
"Chueng-Ryong",
""
]
] |
2506.02846 | Yujin Chen | Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Nie{\ss}ner | PBR-SR: Mesh PBR Texture Super Resolution from 2D Image Priors | Project page: https://terencecyj.github.io/projects/PBR-SR/, Video: https://youtu.be/eaM5S3Mt1RM | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present PBR-SR, a novel method for physically based rendering (PBR) texture super resolution (SR). It outputs high-resolution, high-quality PBR textures from low-resolution (LR) PBR input in a zero-shot manner. PBR-SR leverages an off-the-shelf super-resolution model trained on natural images, and iteratively minimizes the deviations between super-resolution priors and differentiable renderings. These enhancements are then back-projected into the PBR map space in a differentiable manner to produce refined, high-resolution textures. To mitigate view inconsistencies and lighting sensitivity, which is common in view-based super-resolution, our method applies 2D prior constraints across multi-view renderings, iteratively refining the shared, upscaled textures. In parallel, we incorporate identity constraints directly in the PBR texture domain to ensure the upscaled textures remain faithful to the LR input. PBR-SR operates without any additional training or data requirements, relying entirely on pretrained image priors. We demonstrate that our approach produces high-fidelity PBR textures for both artist-designed and AI-generated meshes, outperforming both direct SR models application and prior texture optimization methods. Our results show high-quality outputs in both PBR and rendering evaluations, supporting advanced applications such as relighting. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:15:34 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Chen",
"Yujin",
""
],
[
"Nie",
"Yinyu",
""
],
[
"Ummenhofer",
"Benjamin",
""
],
[
"Birkl",
"Reiner",
""
],
[
"Paulitsch",
"Michael",
""
],
[
"Nießner",
"Matthias",
""
]
] |
2506.02847 | Chunlin Tian | Chunlin Tian, Xinpeng Qin, Kahou Tam, Li Li, Zijian Wang, Yuanzhe Zhao, Minglei Zhang, Chengzhong Xu | CLONE: Customizing LLMs for Efficient Latency-Aware Inference at the Edge | Accepted by USENIX ATC 2025 | null | null | null | cs.AR cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deploying large language models (LLMs) on edge devices is crucial for delivering fast responses and ensuring data privacy. However, the limited storage, weight, and power of edge devices make it difficult to deploy LLM-powered applications. These devices must balance latency requirements with energy consumption and model accuracy. In this paper, we first quantify the challenges of deploying LLMs on off-the-shelf edge devices and then we present CLONE, an in-depth algorithm-hardware co-design at both the model- and system-level that intelligently integrates real-time, energy optimization while maintaining robust generality. In order to maximize the synergistic benefits of these algorithms in always-on and intermediate edge computing settings, we specialize in a 28nm scalable hardware accelerator system. We implement and extensively evaluate CLONE on two off-the-shelf edge platforms. Experiments show that CLONE effectively accelerates the inference process up to 11.92x, and saves energy up to 7.36x, while maintaining high-generation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:16:00 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Tian",
"Chunlin",
""
],
[
"Qin",
"Xinpeng",
""
],
[
"Tam",
"Kahou",
""
],
[
"Li",
"Li",
""
],
[
"Wang",
"Zijian",
""
],
[
"Zhao",
"Yuanzhe",
""
],
[
"Zhang",
"Minglei",
""
],
[
"Xu",
"Chengzhong",
""
]
] |
2506.02848 | Denys Malyshev | Denys Malyshev, Lidiia Zadorozhna, Yuriy Bidasyuk, Andrea Santangelo, Oleg Ruchayskiy | Active galactic nuclei through the prism of galaxy clusters: bounds on axion-like particles | null | null | null | null | astro-ph.HE astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hypothetical axion-like particles (ALPs) are of interest because of their potential to act as dark matter or to reveal information about yet undiscovered fundamental constituents of matter. Such particles can be created when photons traverse regions of magnetic fields. The conversion probability depends on both the magnetic field parameters and photon energy, leading to multiple spectral absorption features as light passes through magnetized regions.
Traditionally, astrophysical searches have focused on detecting such features in individual objects. However, the limited understanding of properties of cosmic magnetic fields have hindered the progress. Here we introduce a new approach by analyzing stacked (rather than individual) spectra of active galactic nuclei (AGNs) positioned behind galaxy clusters -- gigantic magnetic field reservoirs. Stacking efficiently averages over the uncertainties in magnetic fields, revealing a unique step-like spectral signature of photon-to-ALP conversion.
With this approach we advance into previously inaccessible regions of the ALP parameter space for nano-electronvolt masses. Adopting this method will significantly improve existing bounds across a wide range of masses by using different telescopes and increasing the size of the stacked datasets. The Cherenkov Telescope Array Observatory, in particular, will extensively probe the parameter space where ALPs could serve as dark matter. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:19:12 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Malyshev",
"Denys",
""
],
[
"Zadorozhna",
"Lidiia",
""
],
[
"Bidasyuk",
"Yuriy",
""
],
[
"Santangelo",
"Andrea",
""
],
[
"Ruchayskiy",
"Oleg",
""
]
] |
2506.02849 | Alejandro Sanchez Roncero | Alejandro Sanchez Roncero, Olov Andersson, Petter Ogren | Learned Controllers for Agile Quadrotors in Pursuit-Evasion Games | null | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing proliferation of small UAVs in civilian and military airspace has raised critical safety and security concerns, especially when unauthorized or malicious drones enter restricted zones. In this work, we present a reinforcement learning (RL) framework for agile 1v1 quadrotor pursuit-evasion. We train neural network policies to command body rates and collective thrust, enabling high-speed pursuit and evasive maneuvers that fully exploit the quadrotor's nonlinear dynamics. To mitigate nonstationarity and catastrophic forgetting during adversarial co-training, we introduce an Asynchronous Multi-Stage Population-Based (AMSPB) algorithm where, at each stage, either the pursuer or evader learns against a sampled opponent drawn from a growing population of past and current policies. This continual learning setup ensures monotonic performance improvement and retention of earlier strategies. Our results show that (i) rate-based policies achieve significantly higher capture rates and peak speeds than velocity-level baselines, and (ii) AMSPB yields stable, monotonic gains against a suite of benchmark opponents. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:19:23 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Roncero",
"Alejandro Sanchez",
""
],
[
"Andersson",
"Olov",
""
],
[
"Ogren",
"Petter",
""
]
] |
2506.02850 | Mengyue Wang | Mengyue Wang, Shuo Chen, Kristian Kersting, Volker Tresp, Yunpu Ma | METok: Multi-Stage Event-based Token Compression for Efficient Long Video Understanding | 14 pages, 10 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in Video Large Language Models (VLLMs) have significantly enhanced their ability to understand video content. Nonetheless, processing long videos remains challenging due to high computational demands and the redundancy present in the visual data. In this work, we propose METok, a training-free, Multi-stage Event-based Token compression framework designed to accelerate VLLMs' inference while preserving accuracy. METok progressively eliminates redundant visual tokens across three critical stages: (1) event-aware compression during vision encoding, (2) hierarchical token pruning in the prefilling stage based on semantic alignment and event importance, and (3) a decoding-stage KV Cache optimization that further reduces memory consumption. Our experiments on diverse video benchmarks demonstrate that METok achieves an optimal trade-off between efficiency and accuracy by dynamically selecting informative visual tokens. For instance, equipping LongVA-7B with METok realizes an 80.6% FLOPs reduction and 93.5% KV Cache memory savings, all while maintaining comparable or even superior accuracy. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:19:41 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wang",
"Mengyue",
""
],
[
"Chen",
"Shuo",
""
],
[
"Kersting",
"Kristian",
""
],
[
"Tresp",
"Volker",
""
],
[
"Ma",
"Yunpu",
""
]
] |
2506.02851 | Juho Lintuvuori | Zaiyi Shen, Dongfang Fu and Juho Lintuvuori | Inertia-driven propulsion of asymmetric spinner-dimers at moderate Reynolds numbers | 9 pages, 7 figures | Soft Matter 21, 4021-4028 (2025) | 10.1039/D5SM00108K | null | cond-mat.soft physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the translational motion of rotating colloidal systems at moderate Reynolds numbers (Re), focusing on particle dimers in snowman-like configurations under three scenarios: (i) two co-rotating spheres driven by an external field, (ii) two counter-rotating spheres driven by an internal torque as a swimmer, and (iii) a single rotating spinner with a passive sphere for cargo delivery, using hydrodynamic simulations. In all the three cases, the particles are bound together hydrodynamically, and the purely rotational motion of the spinners produces a net propulsion of the dimers along the axis of rotation due to a symmetry breaking. We demonstrate tunable dynamics, where the propulsion direction of the co-rotating dimer can be reversed by tuning the aspect ratio and Reynolds number, as well as cargo transport where a dimer consisting of a single spinner and a passive cargo particle can have a sustained locomotion due to broken head-to-tail symmetry of the overall flow fields. These findings highlight the critical role of inertia in creating locomotion from rotational motion and offer new avenues for controlling and optimizing translational motion in colloidal assemblies through rotational degrees of freedom. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:19:54 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Shen",
"Zaiyi",
""
],
[
"Fu",
"Dongfang",
""
],
[
"Lintuvuori",
"Juho",
""
]
] |
2506.02852 | Yixin Tao | Yun Kuen Cheung, Richard Cole, Yixin Tao | Proportional Response Dynamics in Gross Substitutes Markets | null | null | null | null | cs.GT econ.TH | http://creativecommons.org/licenses/by/4.0/ | Proportional response is a well-established distributed algorithm which has been shown to converge to competitive equilibria in both Fisher and Arrow-Debreu markets, for various sub-families of homogeneous utilities, including linear and constant elasticity of substitution utilities. We propose a natural generalization of proportional response for gross substitutes utilities, and prove that it converges to competitive equilibria in Fisher markets. This is the first convergence result of a proportional response style dynamics in Fisher markets for utilities beyond the homogeneous utilities covered by the Eisenberg-Gale convex program. We show an empirical convergence rate of $O(1/T)$ for the prices. Furthermore, we show that the allocations of a lazy version of the generalized proportional response dynamics converge to competitive equilibria in Arrow-Debreu markets. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:20:27 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Cheung",
"Yun Kuen",
""
],
[
"Cole",
"Richard",
""
],
[
"Tao",
"Yixin",
""
]
] |
2506.02853 | Mingjie Wei | Mingjie Wei, Xuemei Xie, Yutong Zhong, Guangming Shi | Learning Pyramid-structured Long-range Dependencies for 3D Human Pose Estimation | Accepted by IEEE Transactions on Multimedia (TMM) | null | 10.1109/TMM.2025.3535349 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Action coordination in human structure is indispensable for the spatial constraints of 2D joints to recover 3D pose. Usually, action coordination is represented as a long-range dependence among body parts. However, there are two main challenges in modeling long-range dependencies. First, joints should not only be constrained by other individual joints but also be modulated by the body parts. Second, existing methods make networks deeper to learn dependencies between non-linked parts. They introduce uncorrelated noise and increase the model size. In this paper, we utilize a pyramid structure to better learn potential long-range dependencies. It can capture the correlation across joints and groups, which complements the context of the human sub-structure. In an effective cross-scale way, it captures the pyramid-structured long-range dependence. Specifically, we propose a novel Pyramid Graph Attention (PGA) module to capture long-range cross-scale dependencies. It concatenates information from various scales into a compact sequence, and then computes the correlation between scales in parallel. Combining PGA with graph convolution modules, we develop a Pyramid Graph Transformer (PGFormer) for 3D human pose estimation, which is a lightweight multi-scale transformer architecture. It encapsulates human sub-structures into self-attention by pooling. Extensive experiments show that our approach achieves lower error and smaller model size than state-of-the-art methods on Human3.6M and MPI-INF-3DHP datasets. The code is available at https://github.com/MingjieWe/PGFormer. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:21:37 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wei",
"Mingjie",
""
],
[
"Xie",
"Xuemei",
""
],
[
"Zhong",
"Yutong",
""
],
[
"Shi",
"Guangming",
""
]
] |
2506.02854 | Mengmeng Zhang | Mengmeng Zhang, Xingyuan Dai, Yicheng Sun, Jing Wang, Yueyang Yao, Xiaoyan Gong, Fuze Cong, Feiyue Wang, Yisheng Lv | Hierarchical Self-Prompting SAM: A Prompt-Free Medical Image Segmentation Framework | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Although the Segment Anything Model (SAM) is highly effective in natural image segmentation, it requires dependencies on prompts, which limits its applicability to medical imaging where manual prompts are often unavailable. Existing efforts to fine-tune SAM for medical segmentation typically struggle to remove this dependency. We propose Hierarchical Self-Prompting SAM (HSP-SAM), a novel self-prompting framework that enables SAM to achieve strong performance in prompt-free medical image segmentation. Unlike previous self-prompting methods that remain limited to positional prompts similar to vanilla SAM, we are the first to introduce learning abstract prompts during the self-prompting process. This simple and intuitive self-prompting framework achieves superior performance on classic segmentation tasks such as polyp and skin lesion segmentation, while maintaining robustness across diverse medical imaging modalities. Furthermore, it exhibits strong generalization to unseen datasets, achieving improvements of up to 14.04% over previous state-of-the-art methods on some challenging benchmarks. These results suggest that abstract prompts encapsulate richer and higher-dimensional semantic information compared to positional prompts, thereby enhancing the model's robustness and generalization performance. All models and codes will be released upon acceptance. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:23:33 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Zhang",
"Mengmeng",
""
],
[
"Dai",
"Xingyuan",
""
],
[
"Sun",
"Yicheng",
""
],
[
"Wang",
"Jing",
""
],
[
"Yao",
"Yueyang",
""
],
[
"Gong",
"Xiaoyan",
""
],
[
"Cong",
"Fuze",
""
],
[
"Wang",
"Feiyue",
""
],
[
"Lv",
"Yisheng",
""
]
] |
2506.02855 | Yonghui Xia | Lu Weijie, Xia Yonghui | Pugh's global linearization for the nonautonomous unbounded system with $\mu$-dichotomy via Lyapunov theory | 35pages | null | null | null | math.DS math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The classical global linearization theorem for autonomous system given in [C. Pugh, Amer. J. Math., 91 (1969) 363-367] requires that nonlinear system with hyperbolicity satisfies boundedness and Lipschitz continuity.In this paper, we establish an {\em unbounded} global linearization theorem for nonautonomous systems subject to unbounded Lipschitz perturbations, under the assumption that the linear system admits a nonuniform $\mu$-dichotomy (more general than classical exponential dichotomy). To this end, we first develop a comprehensive Lyapunov function framework for systems exhibiting nonuniform $\mu$-dichotomy. Subsequently, we establish a characterization of nonuniform $\mu$-dichotomy in terms of strict quadratic Lyapunov functions. Building upon these theoretical foundations, we then employ these Lyapunov functions to derive a linearization result under the nonuniform $\mu$-dichotomy assumption. In the proof, we give a splitting lemma for nonuniform $\mu$-dichotomy to decouple hyperbolic system into a contractive system and an expansive system. Then we construct a transformation to linearize contractive/expansive system, which is defined by the crossing time with respect to the unit sphere. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:23:35 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Weijie",
"Lu",
""
],
[
"Yonghui",
"Xia",
""
]
] |
2506.02856 | Kimaya Lecamwasam | Kimaya Lecamwasam and Tishya Ray Chaudhuri | Exploring listeners' perceptions of AI-generated and human-composed music for functional emotional applications | 18 pages (includes references and appendix), 2 figures, 5 tables (includes 1 in appendix), and appendix | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | This work investigates how listeners perceive and evaluate AI-generated as compared to human-composed music in the context of emotional resonance and regulation. Across a mixed-methods design, participants were exposed to both AI and human music under various labeling conditions (music correctly labeled as AI- or human-origin, music incorrectly labeled as AI- or human-origin, and unlabeled music) and emotion cases (Calm and Upbeat), and were asked to rate preference, efficacy of target emotion elicitation, and emotional impact. Participants were significantly more likely to rate human-composed music, regardless of labeling, as more effective at eliciting target emotional states, though quantitative analyses revealed no significant differences in emotional response. However, participants were significantly more likely to indicate preference for AI-generated music, yielding further questions regarding the impact of emotional authenticity and perceived authorship on musical appraisal. Qualitative data underscored this, with participants associating humanness with qualities such as imperfection, flow, and 'soul.' These findings challenge the assumption that preference alone signals success in generative music systems. Rather than positioning AI tools as replacements for human creativity or emotional expression, they point toward a more careful design ethos that acknowledges the limits of replication and prioritizes human values such as authenticity, individuality, and emotion regulation in wellness and affective technologies. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:23:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Lecamwasam",
"Kimaya",
""
],
[
"Chaudhuri",
"Tishya Ray",
""
]
] |
2506.02857 | Irene Amerini | Luca Maiano, Fabrizio Casadei, Irene Amerini | Enhancing Abnormality Identification: Robust Out-of-Distribution Strategies for Deepfake Detection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Detecting deepfakes has become a critical challenge in Computer Vision and Artificial Intelligence. Despite significant progress in detection techniques, generalizing them to open-set scenarios continues to be a persistent difficulty. Neural networks are often trained on the closed-world assumption, but with new generative models constantly evolving, it is inevitable to encounter data generated by models that are not part of the training distribution. To address these challenges, in this paper, we propose two novel Out-Of-Distribution (OOD) detection approaches. The first approach is trained to reconstruct the input image, while the second incorporates an attention mechanism for detecting OODs. Our experiments validate the effectiveness of the proposed approaches compared to existing state-of-the-art techniques. Our method achieves promising results in deepfake detection and ranks among the top-performing configurations on the benchmark, demonstrating their potential for robust, adaptable solutions in dynamic, real-world applications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:24:33 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Maiano",
"Luca",
""
],
[
"Casadei",
"Fabrizio",
""
],
[
"Amerini",
"Irene",
""
]
] |
2506.02859 | Asaf Shabtai | Parth Atulbhai Gandhi, Akansha Shukla, David Tayouri, Beni Ifland, Yuval Elovici, Rami Puzis, Asaf Shabtai | ATAG: AI-Agent Application Threat Assessment with Attack Graphs | null | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Evaluating the security of multi-agent systems (MASs) powered by large language models (LLMs) is challenging, primarily because of the systems' complex internal dynamics and the evolving nature of LLM vulnerabilities. Traditional attack graph (AG) methods often lack the specific capabilities to model attacks on LLMs. This paper introduces AI-agent application Threat assessment with Attack Graphs (ATAG), a novel framework designed to systematically analyze the security risks associated with AI-agent applications. ATAG extends the MulVAL logic-based AG generation tool with custom facts and interaction rules to accurately represent AI-agent topologies, vulnerabilities, and attack scenarios. As part of this research, we also created the LLM vulnerability database (LVD) to initiate the process of standardizing LLM vulnerabilities documentation. To demonstrate ATAG's efficacy, we applied it to two multi-agent applications. Our case studies demonstrated the framework's ability to model and generate AGs for sophisticated, multi-step attack scenarios exploiting vulnerabilities such as prompt injection, excessive agency, sensitive information disclosure, and insecure output handling across interconnected agents. ATAG is an important step toward a robust methodology and toolset to help understand, visualize, and prioritize complex attack paths in multi-agent AI systems (MAASs). It facilitates proactive identification and mitigation of AI-agent threats in multi-agent applications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:25:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Gandhi",
"Parth Atulbhai",
""
],
[
"Shukla",
"Akansha",
""
],
[
"Tayouri",
"David",
""
],
[
"Ifland",
"Beni",
""
],
[
"Elovici",
"Yuval",
""
],
[
"Puzis",
"Rami",
""
],
[
"Shabtai",
"Asaf",
""
]
] |
2506.02860 | Wenjing Tang | Wenjing Tang, Xinyu He, Yongxi Huang, Yunxiao Xiao, Cewu Lu, Panpan Cai | Tru-POMDP: Task Planning Under Uncertainty via Tree of Hypotheses and Open-Ended POMDPs | null | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Task planning under uncertainty is essential for home-service robots operating in the real world. Tasks involve ambiguous human instructions, hidden or unknown object locations, and open-vocabulary object types, leading to significant open-ended uncertainty and a boundlessly large planning space. To address these challenges, we propose Tru-POMDP, a planner that combines structured belief generation using Large Language Models (LLMs) with principled POMDP planning. Tru-POMDP introduces a hierarchical Tree of Hypotheses (TOH), which systematically queries an LLM to construct high-quality particle beliefs over possible world states and human goals. We further formulate an open-ended POMDP model that enables rigorous Bayesian belief tracking and efficient belief-space planning over these LLM-generated hypotheses. Experiments on complex object rearrangement tasks across diverse kitchen environments show that Tru-POMDP significantly outperforms state-of-the-art LLM-based and LLM-tree-search hybrid planners, achieving higher success rates with significantly better plans, stronger robustness to ambiguity and occlusion, and greater planning efficiency. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:26:08 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Tang",
"Wenjing",
""
],
[
"He",
"Xinyu",
""
],
[
"Huang",
"Yongxi",
""
],
[
"Xiao",
"Yunxiao",
""
],
[
"Lu",
"Cewu",
""
],
[
"Cai",
"Panpan",
""
]
] |
2506.02861 | Robbe De Mey | Robbe de Mey, Spencer W. Jolly, Alexandre Locquet, Martin Virte | Agile laser wavelength tuning using dynamic targeting | 4 pages, 4 figures | null | null | null | physics.optics nlin.CD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tunable lasers are essential and versatile tools in photonics, with applications spanning telecommunications, spectroscopy, and sensing. Advancements have aimed to expand tuning ranges, suppress mode hopping, and enable photonic integration. In this work, we explore the adaptation of dynamic targeting, a technique originally developed to stabilize lasers under optical feedback, as a method for achieving agile, fast, and continuous wavelength tuning. By adjusting the feedback rate and phase, we enable a stable and controlled frequency shift. We experimentally demonstrate reliable and reproducible tuning over 2.1 GHz using a free-space optical setup. Simulations further suggest that this approach could extend the tuning range to tens of GHz, with a potential scan speed exceeding $10^{17}$ Hz/s. These results highlight dynamic targeting as a promising route toward agile frequency control in semiconductor lasers. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:26:31 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"de Mey",
"Robbe",
""
],
[
"Jolly",
"Spencer W.",
""
],
[
"Locquet",
"Alexandre",
""
],
[
"Virte",
"Martin",
""
]
] |
2506.02862 | Lukas K\"ohldorfer | Peter Balazs, Lukas K\"ohldorfer and Michael Speckbacher | Localized frames without inequalities | null | null | null | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | We consider countable families of vectors in a separable Hilbert space, which are mutually localized with respect to a fixed localized Riesz basis. We prove the equivalence of the frame property and nine conditions that do not involve any inequalities. This is done by studying the properties of their frame-related operators on the co-orbit spaces generated by the reference Riesz basis. We apply our main result to the setting of shift-invariant spaces and obtain new conditions for stable sets of sampling. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:27:48 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Balazs",
"Peter",
""
],
[
"Köhldorfer",
"Lukas",
""
],
[
"Speckbacher",
"Michael",
""
]
] |
2506.02863 | Helin Wang | Helin Wang, Jiarui Hai, Dading Chong, Karan Thakkar, Tiantian Feng, Dongchao Yang, Junhyeok Lee, Laureano Moro Velazquez, Jesus Villalba, Zengyi Qin, Shrikanth Narayanan, Mounya Elhiali, Najim Dehak | CapSpeech: Enabling Downstream Applications in Style-Captioned Text-to-Speech | null | null | null | null | eess.AS cs.AI cs.SD | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent advancements in generative artificial intelligence have significantly transformed the field of style-captioned text-to-speech synthesis (CapTTS). However, adapting CapTTS to real-world applications remains challenging due to the lack of standardized, comprehensive datasets and limited research on downstream tasks built upon CapTTS. To address these gaps, we introduce CapSpeech, a new benchmark designed for a series of CapTTS-related tasks, including style-captioned text-to-speech synthesis with sound events (CapTTS-SE), accent-captioned TTS (AccCapTTS), emotion-captioned TTS (EmoCapTTS), and text-to-speech synthesis for chat agent (AgentTTS). CapSpeech comprises over 10 million machine-annotated audio-caption pairs and nearly 0.36 million human-annotated audio-caption pairs. In addition, we introduce two new datasets collected and recorded by a professional voice actor and experienced audio engineers, specifically for the AgentTTS and CapTTS-SE tasks. Alongside the datasets, we conduct comprehensive experiments using both autoregressive and non-autoregressive models on CapSpeech. Our results demonstrate high-fidelity and highly intelligible speech synthesis across a diverse range of speaking styles. To the best of our knowledge, CapSpeech is the largest available dataset offering comprehensive annotations for CapTTS-related tasks. The experiments and findings further provide valuable insights into the challenges of developing CapTTS systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:28:55 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wang",
"Helin",
""
],
[
"Hai",
"Jiarui",
""
],
[
"Chong",
"Dading",
""
],
[
"Thakkar",
"Karan",
""
],
[
"Feng",
"Tiantian",
""
],
[
"Yang",
"Dongchao",
""
],
[
"Lee",
"Junhyeok",
""
],
[
"Velazquez",
"Laureano Moro",
""
],
[
"Villalba",
"Jesus",
""
],
[
"Qin",
"Zengyi",
""
],
[
"Narayanan",
"Shrikanth",
""
],
[
"Elhiali",
"Mounya",
""
],
[
"Dehak",
"Najim",
""
]
] |
2506.02864 | Changyi Xiao | Changyi Xiao, Mengdi Zhang, Yixin Cao | BNPO: Beta Normalization Policy Optimization | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent studies, including DeepSeek-R1 and Kimi-k1.5, have demonstrated that reinforcement learning with rule-based, binary-valued reward functions can significantly enhance the reasoning capabilities of large language models. These models primarily utilize REINFORCE-based policy optimization techniques, such as REINFORCE with baseline and group relative policy optimization (GRPO). However, a key limitation remains: current policy optimization methods either neglect reward normalization or employ static normalization strategies, which fail to adapt to the dynamic nature of policy updates during training. This may result in unstable gradient estimates and hinder training stability. To address this issue, we propose Beta Normalization Policy Optimization (BNPO), a novel policy optimization method that adaptively normalizes rewards using a Beta distribution with dynamically updated parameters. BNPO aligns the normalization with the changing policy distribution, enabling more precise and lower-variance gradient estimation, which in turn promotes stable training dynamics. We provide theoretical analysis demonstrating BNPO's variance-reducing properties and show that it generalizes both REINFORCE and GRPO under binary-valued reward settings. Furthermore, we introduce an advantage decomposition mechanism to extend BNPO's applicability to more complex reward systems. Experimental results confirm that BNPO achieves state-of-the-art performance among policy optimization methods on reasoning tasks. The code is available at https://github.com/changyi7231/BNPO. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:28:57 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Xiao",
"Changyi",
""
],
[
"Zhang",
"Mengdi",
""
],
[
"Cao",
"Yixin",
""
]
] |
2506.02865 | Maxime Theillard | Mathieu Andreux, Breno Baldas Skuk, Hamza Benchekroun, Emilien Bir\'e, Antoine Bonnet, Riaz Bordie, Matthias Brunel, Pierre-Louis Cedoz, Antoine Chassang, Micka\"el Chen, Alexandra D. Constantinou, Antoine d'Andign\'e, Hubert de La Jonqui\`ere, Aur\'elien Delfosse, Ludovic Denoyer, Alexis Deprez, Augustin Derupti, Michael Eickenberg, Math\"is Federico, Charles Kantor, Xavier Koegler, Yann Labb\'e, Matthew C. H. Lee, Erwan Le Jumeau de Kergaradec, Amir Mahla, Avshalom Manevich, Adrien Maret, Charles Masson, Rafa\"el Maurin, Arturo Mena, Philippe Modard, Axel Moyal, Axel Nguyen Kerbel, Julien Revelle, Mats L. Richter, Mar\'ia Santos, Laurent Sifre, Maxime Theillard, Marc Thibault, Louis Thiry, L\'eo Tronchon, Nicolas Usunier, Tony Wu | Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights | Alphabetical order | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | We present Surfer-H, a cost-efficient web agent that integrates Vision-Language Models (VLM) to perform user-defined tasks on the web. We pair it with Holo1, a new open-weight collection of VLMs specialized in web navigation and information extraction. Holo1 was trained on carefully curated data sources, including open-access web content, synthetic examples, and self-produced agentic data. Holo1 tops generalist User Interface (UI) benchmarks as well as our new web UI localization benchmark, WebClick. When powered by Holo1, Surfer-H achieves a 92.2% state-of-the-art performance on WebVoyager, striking a Pareto-optimal balance between accuracy and cost-efficiency. To accelerate research advancement in agentic systems, we are open-sourcing both our WebClick evaluation dataset and the Holo1 model weights. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:29:03 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Andreux",
"Mathieu",
""
],
[
"Skuk",
"Breno Baldas",
""
],
[
"Benchekroun",
"Hamza",
""
],
[
"Biré",
"Emilien",
""
],
[
"Bonnet",
"Antoine",
""
],
[
"Bordie",
"Riaz",
""
],
[
"Brunel",
"Matthias",
""
],
[
"Cedoz",
"Pierre-Louis",
""
],
[
"Chassang",
"Antoine",
""
],
[
"Chen",
"Mickaël",
""
],
[
"Constantinou",
"Alexandra D.",
""
],
[
"d'Andigné",
"Antoine",
""
],
[
"de La Jonquière",
"Hubert",
""
],
[
"Delfosse",
"Aurélien",
""
],
[
"Denoyer",
"Ludovic",
""
],
[
"Deprez",
"Alexis",
""
],
[
"Derupti",
"Augustin",
""
],
[
"Eickenberg",
"Michael",
""
],
[
"Federico",
"Mathïs",
""
],
[
"Kantor",
"Charles",
""
],
[
"Koegler",
"Xavier",
""
],
[
"Labbé",
"Yann",
""
],
[
"Lee",
"Matthew C. H.",
""
],
[
"de Kergaradec",
"Erwan Le Jumeau",
""
],
[
"Mahla",
"Amir",
""
],
[
"Manevich",
"Avshalom",
""
],
[
"Maret",
"Adrien",
""
],
[
"Masson",
"Charles",
""
],
[
"Maurin",
"Rafaël",
""
],
[
"Mena",
"Arturo",
""
],
[
"Modard",
"Philippe",
""
],
[
"Moyal",
"Axel",
""
],
[
"Kerbel",
"Axel Nguyen",
""
],
[
"Revelle",
"Julien",
""
],
[
"Richter",
"Mats L.",
""
],
[
"Santos",
"María",
""
],
[
"Sifre",
"Laurent",
""
],
[
"Theillard",
"Maxime",
""
],
[
"Thibault",
"Marc",
""
],
[
"Thiry",
"Louis",
""
],
[
"Tronchon",
"Léo",
""
],
[
"Usunier",
"Nicolas",
""
],
[
"Wu",
"Tony",
""
]
] |
2506.02866 | Ahsan Baidar Bakht | Ahsan Baidar Bakht, Muhayy Ud Din, Sajid Javed, Irfan Hussain | MVTD: A Benchmark Dataset for Maritime Visual Object Tracking | Submited to Nature Scientific Data | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Visual Object Tracking (VOT) is a fundamental task with widespread applications in autonomous navigation, surveillance, and maritime robotics. Despite significant advances in generic object tracking, maritime environments continue to present unique challenges, including specular water reflections, low-contrast targets, dynamically changing backgrounds, and frequent occlusions. These complexities significantly degrade the performance of state-of-the-art tracking algorithms, highlighting the need for domain-specific datasets. To address this gap, we introduce the Maritime Visual Tracking Dataset (MVTD), a comprehensive and publicly available benchmark specifically designed for maritime VOT. MVTD comprises 182 high-resolution video sequences, totaling approximately 150,000 frames, and includes four representative object classes: boat, ship, sailboat, and unmanned surface vehicle (USV). The dataset captures a diverse range of operational conditions and maritime scenarios, reflecting the real-world complexities of maritime environments. We evaluated 14 recent SOTA tracking algorithms on the MVTD benchmark and observed substantial performance degradation compared to their performance on general-purpose datasets. However, when fine-tuned on MVTD, these models demonstrate significant performance gains, underscoring the effectiveness of domain adaptation and the importance of transfer learning in specialized tracking contexts. The MVTD dataset fills a critical gap in the visual tracking community by providing a realistic and challenging benchmark for maritime scenarios. Dataset and Source Code can be accessed here "https://github.com/AhsanBaidar/MVTD". | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:30:11 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Bakht",
"Ahsan Baidar",
""
],
[
"Din",
"Muhayy Ud",
""
],
[
"Javed",
"Sajid",
""
],
[
"Hussain",
"Irfan",
""
]
] |
2506.02868 | Amal Perera | Amal S. Perera, David Fernandez, Chandi Witharana, Elias Manos, Michael Pimenta, Anna K. Liljedahl, Ingmar Nitze, Yili Yang, Todd Nicholson, Chia-Yu Hsu, Wenwen Li, Guido Grosse | Pan-Arctic Permafrost Landform and Human-built Infrastructure Feature Detection with Vision Transformers and Location Embeddings | 20 pages, 2 column IEEE format, 13 Figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Accurate mapping of permafrost landforms, thaw disturbances, and human-built infrastructure at pan-Arctic scale using sub-meter satellite imagery is increasingly critical. Handling petabyte-scale image data requires high-performance computing and robust feature detection models. While convolutional neural network (CNN)-based deep learning approaches are widely used for remote sensing (RS),similar to the success in transformer based large language models, Vision Transformers (ViTs) offer advantages in capturing long-range dependencies and global context via attention mechanisms. ViTs support pretraining via self-supervised learning-addressing the common limitation of labeled data in Arctic feature detection and outperform CNNs on benchmark datasets. Arctic also poses challenges for model generalization, especially when features with the same semantic class exhibit diverse spectral characteristics. To address these issues for Arctic feature detection, we integrate geospatial location embeddings into ViTs to improve adaptation across regions. This work investigates: (1) the suitability of pre-trained ViTs as feature extractors for high-resolution Arctic remote sensing tasks, and (2) the benefit of combining image and location embeddings. Using previously published datasets for Arctic feature detection, we evaluate our models on three tasks-detecting ice-wedge polygons (IWP), retrogressive thaw slumps (RTS), and human-built infrastructure. We empirically explore multiple configurations to fuse image embeddings and location embeddings. Results show that ViTs with location embeddings outperform prior CNN-based models on two of the three tasks including F1 score increase from 0.84 to 0.92 for RTS detection, demonstrating the potential of transformer-based models with spatial awareness for Arctic RS applications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:34:01 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Perera",
"Amal S.",
""
],
[
"Fernandez",
"David",
""
],
[
"Witharana",
"Chandi",
""
],
[
"Manos",
"Elias",
""
],
[
"Pimenta",
"Michael",
""
],
[
"Liljedahl",
"Anna K.",
""
],
[
"Nitze",
"Ingmar",
""
],
[
"Yang",
"Yili",
""
],
[
"Nicholson",
"Todd",
""
],
[
"Hsu",
"Chia-Yu",
""
],
[
"Li",
"Wenwen",
""
],
[
"Grosse",
"Guido",
""
]
] |
2506.02869 | Leonardo Baggiani | Leonardo Baggiani, Martin Herdegen, Leandro S\'anchez-Betancourt | Optimal Dynamic Fees in Automated Market Makers | 18 pages | null | null | null | q-fin.TR q-fin.MF | http://creativecommons.org/licenses/by/4.0/ | Automated Market Makers (AMMs) are emerging as a popular decentralised trading platform. In this work, we determine the optimal dynamic fees in a constant function market maker. We find approximate closed-form solutions to the control problem and study the optimal fee structure. We find that there are two distinct fee regimes: one in which the AMM imposes higher fees to deter arbitrageurs, and another where fees are lowered to increase volatility and attract noise traders. Our results also show that dynamic fees that are linear in inventory and are sensitive to changes in the external price are a good approximation of the optimal fee structure and thus constitute suitable candidates when designing fees for AMMs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:34:28 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Baggiani",
"Leonardo",
""
],
[
"Herdegen",
"Martin",
""
],
[
"Sánchez-Betancourt",
"Leandro",
""
]
] |
2506.02870 | Jianli Xiang | Jianli Xiang | Uniqueness in determining a convex polygonal source of an elastic body | null | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we consider the time-harmonic inverse elastic source problem of a fixed frequency for the Navier equation in two dimensions. We show that a convex polygon can be uniquely determined by a single far field measurement. Our approach relies on the corner singularity analysis of solutions to the inhomogeneous Navier equation with a source term in a sector. This paper also contributes to corner scattering theory for the Navier equation in an non-convex domain. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:35:35 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Xiang",
"Jianli",
""
]
] |
2506.02871 | Luca Vai | Indranil Biswas, Alessandro Ghigi, Luca Vai | Prym varieties and projective structures on Riemann surfaces | null | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given an \'etale double covering $\pi\, :\, \widetilde{C}\, \longrightarrow\, C$ of compact Riemannsurfaces with $C$ of genus at least two, we use the Prym variety of the cover to construct canonical projective structures on both $\widetilde C$ and $C$. This construction can be interpreted as a section of an affine bundle over the moduli space of \'etale double covers. The $\overline{\partial}$--derivative of this section is a (1,1)--form on the moduli space. We compute this derivative in terms of Thetanullwert maps. Using the Schottky--Jung identities we show that, in general, the projective structure on $C$ depends on the cover. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:37:42 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Biswas",
"Indranil",
""
],
[
"Ghigi",
"Alessandro",
""
],
[
"Vai",
"Luca",
""
]
] |
2506.02872 | Ludovic Moncla | Ludovic Moncla and H\'edi Zeghidi | Token and Span Classification for Entity Recognition in French Historical Encyclopedias | null | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | Named Entity Recognition (NER) in historical texts presents unique challenges due to non-standardized language, archaic orthography, and nested or overlapping entities. This study benchmarks a diverse set of NER approaches, ranging from classical Conditional Random Fields (CRFs) and spaCy-based models to transformer-based architectures such as CamemBERT and sequence-labeling models like Flair. Experiments are conducted on the GeoEDdA dataset, a richly annotated corpus derived from 18th-century French encyclopedias. We propose framing NER as both token-level and span-level classification to accommodate complex nested entity structures typical of historical documents. Additionally, we evaluate the emerging potential of few-shot prompting with generative language models for low-resource scenarios. Our results demonstrate that while transformer-based models achieve state-of-the-art performance, especially on nested entities, generative models offer promising alternatives when labeled data are scarce. The study highlights ongoing challenges in historical NER and suggests avenues for hybrid approaches combining symbolic and neural methods to better capture the intricacies of early modern French text. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:37:44 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Moncla",
"Ludovic",
""
],
[
"Zeghidi",
"Hédi",
""
]
] |
2506.02873 | Matthew Kowal | Matthew Kowal, Jasper Timm, Jean-Francois Godbout, Thomas Costello, Antonio A. Arechar, Gordon Pennycook, David Rand, Adam Gleave, Kellin Pelrine | It's the Thought that Counts: Evaluating the Attempts of Frontier LLMs to Persuade on Harmful Topics | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Persuasion is a powerful capability of large language models (LLMs) that both enables beneficial applications (e.g. helping people quit smoking) and raises significant risks (e.g. large-scale, targeted political manipulation). Prior work has found models possess a significant and growing persuasive capability, measured by belief changes in simulated or real users. However, these benchmarks overlook a crucial risk factor: the propensity of a model to attempt to persuade in harmful contexts. Understanding whether a model will blindly ``follow orders'' to persuade on harmful topics (e.g. glorifying joining a terrorist group) is key to understanding the efficacy of safety guardrails. Moreover, understanding if and when a model will engage in persuasive behavior in pursuit of some goal is essential to understanding the risks from agentic AI systems. We propose the Attempt to Persuade Eval (APE) benchmark, that shifts the focus from persuasion success to persuasion attempts, operationalized as a model's willingness to generate content aimed at shaping beliefs or behavior. Our evaluation framework probes frontier LLMs using a multi-turn conversational setup between simulated persuader and persuadee agents. APE explores a diverse spectrum of topics including conspiracies, controversial issues, and non-controversially harmful content. We introduce an automated evaluator model to identify willingness to persuade and measure the frequency and context of persuasive attempts. We find that many open and closed-weight models are frequently willing to attempt persuasion on harmful topics and that jailbreaking can increase willingness to engage in such behavior. Our results highlight gaps in current safety guardrails and underscore the importance of evaluating willingness to persuade as a key dimension of LLM risk. APE is available at github.com/AlignmentResearch/AttemptPersuadeEval | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:37:51 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kowal",
"Matthew",
""
],
[
"Timm",
"Jasper",
""
],
[
"Godbout",
"Jean-Francois",
""
],
[
"Costello",
"Thomas",
""
],
[
"Arechar",
"Antonio A.",
""
],
[
"Pennycook",
"Gordon",
""
],
[
"Rand",
"David",
""
],
[
"Gleave",
"Adam",
""
],
[
"Pelrine",
"Kellin",
""
]
] |
2506.02874 | Yonghui Xia | Lu Weijie, Piccione Paolo, Xia Yonghui | Stable invariant manifold for generalized ODEs with applications to measure differential equations | 23 pages. arXiv admin note: text overlap with arXiv:2301.09955 by other authors | null | null | null | math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper establishes the stable invariant manifold for a new kind of differential equations defined by Kurzweil integral, so-called {\em generalized ODEs} on a Banach space. The nonlinear generalized ODEs are formulated as $$ \frac{dz}{d\tau}=D[\Lambda(t)z+F(z,t)], $$ where $\Lambda(t)$ is a bounded linear operator on a Banach space $\mathscr{Z}$ and $F(z,t)$ is a nonlinear Kurzweil integrable function on $\mathscr{Z}$. The letter $D$ represents that generalized ODEs are defined via its solution, and $\frac{dz}{d\tau}$ only a notation. Hence, generalized ODEs are fundamentally a notational representation of a class of integral equations. Due to the differences between the theory of generalized ODEs and ODEs, it is difficult to extended the stable manifold theorem of ODEs to generalized ODEs. In order to overcome the difficulty, we establish a generalized Lyapunov-Perron equation in the frame of Kurzweil integral theory. Subsequently, we present a stable invariant manifold theorem for nonlinear generalized ODEs when their linear parts exhibit an exponential dichotomy. As effective applications, we finally derive results concerning the existence of stable manifold for measure differential equations and impulsive differential equations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:39:40 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Weijie",
"Lu",
""
],
[
"Paolo",
"Piccione",
""
],
[
"Yonghui",
"Xia",
""
]
] |
2506.02875 | Shushi Wang | Xiaohong Liu, Xiongkuo Min, Qiang Hu, Xiaoyun Zhang, Jie Guo, Guangtao Zhai, Shushi Wang, Yingjie Zhou, Lu Liu, Jingxin Li, Liu Yang, Farong Wen, Li Xu, Yanwei Jiang, Xilei Zhu, Chunyi Li, Zicheng Zhang, Huiyu Duan, Xiele Wu, Yixuan Gao, Yuqin Cao, Jun Jia, Wei Sun, Jiezhang Cao, Radu Timofte, Baojun Li, Jiamian Huang, Dan Luo, Tao Liu, Weixia Zhang, Bingkun Zheng, Junlin Chen, Ruikai Zhou, Meiya Chen, Yu Wang, Hao Jiang, Xiantao Li, Yuxiang Jiang, Jun Tang, Yimeng Zhao, Bo Hu, Zelu Qi, Chaoyang Zhang, Fei Zhao, Ping Shi, Lingzhi Fu, Heng Cong, Shuai He, Rongyu Zhang, Jiarong He, Zongyao Hu, Wei Luo, Zihao Yu, Fengbin Guan, Yiting Lu, Xin Li, Zhibo Chen, Mengjing Su, Yi Wang, Tuo Chen, Chunxiao Li, Shuaiyu Zhao, Jiaxin Wen, Chuyi Lin, Sitong Liu, Ningxin Chu, Jing Wan, Yu Zhou, Baoying Chen, Jishen Zeng, Jiarui Liu, Xianjin Liu, Xin Chen, Lanzhi Zhou, Hangyu Li, You Han, Bibo Xiang, Zhenjie Liu, Jianzhang Lu, Jialin Gui, Renjie Lu, Shangfei Wang, Donghao Zhou, Jingyu Lin, Quanjian Song, Jiancheng Huang, Yufeng Yang, Changwei Wang, Shupeng Zhong, Yang Yang, Lihuo He, Jia Liu, Yuting Xing, Tida Fang, Yuchun Jin | NTIRE 2025 XGC Quality Assessment Challenge: Methods and Results | NTIRE 2025 XGC Quality Assessment Challenge Report. arXiv admin note: text overlap with arXiv:2404.16687 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper reports on the NTIRE 2025 XGC Quality Assessment Challenge, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2025. This challenge is to address a major challenge in the field of video and talking head processing. The challenge is divided into three tracks, including user generated video, AI generated video and talking head. The user-generated video track uses the FineVD-GC, which contains 6,284 user generated videos. The user-generated video track has a total of 125 registered participants. A total of 242 submissions are received in the development phase, and 136 submissions are received in the test phase. Finally, 5 participating teams submitted their models and fact sheets. The AI generated video track uses the Q-Eval-Video, which contains 34,029 AI-Generated Videos (AIGVs) generated by 11 popular Text-to-Video (T2V) models. A total of 133 participants have registered in this track. A total of 396 submissions are received in the development phase, and 226 submissions are received in the test phase. Finally, 6 participating teams submitted their models and fact sheets. The talking head track uses the THQA-NTIRE, which contains 12,247 2D and 3D talking heads. A total of 89 participants have registered in this track. A total of 225 submissions are received in the development phase, and 118 submissions are received in the test phase. Finally, 8 participating teams submitted their models and fact sheets. Each participating team in every track has proposed a method that outperforms the baseline, which has contributed to the development of fields in three tracks. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:39:57 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Liu",
"Xiaohong",
""
],
[
"Min",
"Xiongkuo",
""
],
[
"Hu",
"Qiang",
""
],
[
"Zhang",
"Xiaoyun",
""
],
[
"Guo",
"Jie",
""
],
[
"Zhai",
"Guangtao",
""
],
[
"Wang",
"Shushi",
""
],
[
"Zhou",
"Yingjie",
""
],
[
"Liu",
"Lu",
""
],
[
"Li",
"Jingxin",
""
],
[
"Yang",
"Liu",
""
],
[
"Wen",
"Farong",
""
],
[
"Xu",
"Li",
""
],
[
"Jiang",
"Yanwei",
""
],
[
"Zhu",
"Xilei",
""
],
[
"Li",
"Chunyi",
""
],
[
"Zhang",
"Zicheng",
""
],
[
"Duan",
"Huiyu",
""
],
[
"Wu",
"Xiele",
""
],
[
"Gao",
"Yixuan",
""
],
[
"Cao",
"Yuqin",
""
],
[
"Jia",
"Jun",
""
],
[
"Sun",
"Wei",
""
],
[
"Cao",
"Jiezhang",
""
],
[
"Timofte",
"Radu",
""
],
[
"Li",
"Baojun",
""
],
[
"Huang",
"Jiamian",
""
],
[
"Luo",
"Dan",
""
],
[
"Liu",
"Tao",
""
],
[
"Zhang",
"Weixia",
""
],
[
"Zheng",
"Bingkun",
""
],
[
"Chen",
"Junlin",
""
],
[
"Zhou",
"Ruikai",
""
],
[
"Chen",
"Meiya",
""
],
[
"Wang",
"Yu",
""
],
[
"Jiang",
"Hao",
""
],
[
"Li",
"Xiantao",
""
],
[
"Jiang",
"Yuxiang",
""
],
[
"Tang",
"Jun",
""
],
[
"Zhao",
"Yimeng",
""
],
[
"Hu",
"Bo",
""
],
[
"Qi",
"Zelu",
""
],
[
"Zhang",
"Chaoyang",
""
],
[
"Zhao",
"Fei",
""
],
[
"Shi",
"Ping",
""
],
[
"Fu",
"Lingzhi",
""
],
[
"Cong",
"Heng",
""
],
[
"He",
"Shuai",
""
],
[
"Zhang",
"Rongyu",
""
],
[
"He",
"Jiarong",
""
],
[
"Hu",
"Zongyao",
""
],
[
"Luo",
"Wei",
""
],
[
"Yu",
"Zihao",
""
],
[
"Guan",
"Fengbin",
""
],
[
"Lu",
"Yiting",
""
],
[
"Li",
"Xin",
""
],
[
"Chen",
"Zhibo",
""
],
[
"Su",
"Mengjing",
""
],
[
"Wang",
"Yi",
""
],
[
"Chen",
"Tuo",
""
],
[
"Li",
"Chunxiao",
""
],
[
"Zhao",
"Shuaiyu",
""
],
[
"Wen",
"Jiaxin",
""
],
[
"Lin",
"Chuyi",
""
],
[
"Liu",
"Sitong",
""
],
[
"Chu",
"Ningxin",
""
],
[
"Wan",
"Jing",
""
],
[
"Zhou",
"Yu",
""
],
[
"Chen",
"Baoying",
""
],
[
"Zeng",
"Jishen",
""
],
[
"Liu",
"Jiarui",
""
],
[
"Liu",
"Xianjin",
""
],
[
"Chen",
"Xin",
""
],
[
"Zhou",
"Lanzhi",
""
],
[
"Li",
"Hangyu",
""
],
[
"Han",
"You",
""
],
[
"Xiang",
"Bibo",
""
],
[
"Liu",
"Zhenjie",
""
],
[
"Lu",
"Jianzhang",
""
],
[
"Gui",
"Jialin",
""
],
[
"Lu",
"Renjie",
""
],
[
"Wang",
"Shangfei",
""
],
[
"Zhou",
"Donghao",
""
],
[
"Lin",
"Jingyu",
""
],
[
"Song",
"Quanjian",
""
],
[
"Huang",
"Jiancheng",
""
],
[
"Yang",
"Yufeng",
""
],
[
"Wang",
"Changwei",
""
],
[
"Zhong",
"Shupeng",
""
],
[
"Yang",
"Yang",
""
],
[
"He",
"Lihuo",
""
],
[
"Liu",
"Jia",
""
],
[
"Xing",
"Yuting",
""
],
[
"Fang",
"Tida",
""
],
[
"Jin",
"Yuchun",
""
]
] |
2506.02876 | Victor P. Debattista | Luis M. San Martin Fernandez, Steven Gough-Kelly, Victor P. Debattista, Oscar A. Gonzalez, Ilin Lazar, Alvaro Rojas-Arriagada, Leandro Beraldo e Silva | The chemical and spatial variations of the bulge's velocity ellipsoids | Accepted to MNRAS | null | 10.1093/mnras/staf887 | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We study the velocity ellipsoids in an $N$-body$+$SPH simulation of a barred galaxy which forms a bar with a BP bulge. We focus on the 2D kinematics, and quantify the velocity ellipses by the anisotropy, $\beta_{ij}$, the correlation, $\rho_{ij}$, and the vertex deviation, $l_{\rm v}$. We explore the variations in these quantities based on stellar age within the bulge and compare these results with the Milky Way's bulge using data from APOGEE DR16 and {\it Gaia} DR3. We first explore the variation of the model's velocity ellipses in galactocentric velocities, $v_R$ and $v_\phi$, for two bulge populations, a (relatively) young one and an old one. The bar imprints quadrupoles on the distribution of ellipse properties, which are stronger in the young population, as expected from their stronger bar. The quadrupoles are distorted if we use heliocentric velocities $v_r$ and $v_l$. We then project these kinematics along the line of sight onto the $(l,b)$-plane. Along the minor axis $\beta_{rl}$ changes from positive at low $|b|$ to negative at large $|b|$, crossing over at lower $|b|$ in the young stars. Consequently the vertex deviation peaks at lower $|b|$ in the young population, but reaches similar peak values in the old. The $\rho_{rl}$ is much stronger in the young stars, and traces the bar strength. The APOGEE stars split by the median [Fe/H] follow the same trends. Lastly we explore the velocity ellipses across the entire bulge region in $(l,b)$ space, finding good qualitative agreement between the model and observations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:42:13 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Fernandez",
"Luis M. San Martin",
""
],
[
"Gough-Kelly",
"Steven",
""
],
[
"Debattista",
"Victor P.",
""
],
[
"Gonzalez",
"Oscar A.",
""
],
[
"Lazar",
"Ilin",
""
],
[
"Rojas-Arriagada",
"Alvaro",
""
],
[
"Silva",
"Leandro Beraldo e",
""
]
] |
2506.02877 | Taro Suzuki | Taro Suzuki, Shotaro Kojima, Kazunori Ohno, Naoto Miyamoto, Takahiro Suzuki, Kimitaka Asano, Tomohiro Komatsu, Hiroto Kakizaki | Automatic Operation of an Articulated Dump Truck: State Estimation by Combined QZSS CLAS and Moving-Base RTK Using Multiple GNSS Receivers | Accepted to the ION 2024 Pacific PNT Meeting | Proceedings of the ION 2024 Pacific PNT Meeting, Honolulu, Hawaii, April 2024, pp. 65-74 | 10.33012/2024.19617 | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Labor shortage due to the declining birth rate has become a serious problem in the construction industry, and automation of construction work is attracting attention as a solution to this problem. This paper proposes a method to realize state estimation of dump truck position, orientation and articulation angle using multiple GNSS for automatic operation of dump trucks. RTK-GNSS is commonly used for automation of construction equipment, but in mountainous areas, mobile networks often unstable, and RTK-GNSS using GNSS reference stations cannot be used. Therefore, this paper develops a state estimation method for dump trucks that does not require a GNSS reference station by using the Centimeter Level Augmentation Service (CLAS) of the Japanese Quasi-Zenith Satellite System (QZSS). Although CLAS is capable of centimeter-level position estimation, its positioning accuracy and ambiguity fix rate are lower than those of RTK-GNSS. To solve this problem, we construct a state estimation method by factor graph optimization that combines CLAS positioning and moving-base RTK-GNSS between multiple GNSS antennas. Evaluation tests under real-world environments have shown that the proposed method can estimate the state of dump trucks with the same accuracy as conventional RTK-GNSS, but does not require a GNSS reference station. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:44:12 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Suzuki",
"Taro",
""
],
[
"Kojima",
"Shotaro",
""
],
[
"Ohno",
"Kazunori",
""
],
[
"Miyamoto",
"Naoto",
""
],
[
"Suzuki",
"Takahiro",
""
],
[
"Asano",
"Kimitaka",
""
],
[
"Komatsu",
"Tomohiro",
""
],
[
"Kakizaki",
"Hiroto",
""
]
] |
2506.02878 | Jintian Shao | Jintian Shao, Yiming Cheng | CoT is Not True Reasoning, It Is Just a Tight Constraint to Imitate: A Theory Perspective | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Chain-of-Thought (CoT) prompting has demonstrably enhanced the performance of Large Language Models on tasks requiring multi-step inference. This success has led to widespread claims of emergent reasoning capabilities in these models. In this paper, we present a theoretical counter-perspective: Chain-of-Thought (CoT) does not elicit genuine, abstract reasoning. Instead, we argue that Chain-of-Thought functions as a powerful structural constraint that guides Large Language Models to imitate the form of reasoning. By forcing the generation of intermediate steps, Chain-of-Thought leverages the model immense capacity for sequence prediction and pattern matching, effectively constraining its output to sequences that resemble coherent thought processes. Chain-of-Thought (CoT) prompting has demonstrably enhanced the performance of Large Language Models on tasks requiring multi-step inference. This success has led to widespread claims of emergent reasoning capabilities in these models. In this paper, we present a theoretical counter-perspective: Chain-of-Thought (CoT) does not elicit genuine, abstract reasoning. Instead, we argue that Chain-of-Thought functions as a powerful structural constraint that guides Large Language Models to imitate the form of reasoning. By forcing the generation of intermediate steps, Chain-of-Thought leverages the model immense capacity for sequence prediction and pattern matching, effectively constraining its output to sequences that resemble coherent thought processes. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:45:01 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Shao",
"Jintian",
""
],
[
"Cheng",
"Yiming",
""
]
] |
2506.02879 | Yilong Song | Yilong Song, Peijin Li, Bin Gao, Kun Yuan | Distributed Retraction-Free and Communication-Efficient Optimization on the Stiefel Manifold | null | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | Optimization problems on the Stiefel manifold, ranging from principal component analysis to enhancing neural network robustness, are ubiquitous in machine learning. The Landing algorithm avoids computationally expensive retraction operations on manifolds, making it highly competitive for large-scale problems. This paper extends this method to distributed settings, introducing *EF-Landing*, the first retraction-free and communication-efficient algorithm for distributed stochastic optimization on the Stiefel manifold. By incorporating communication compression and error feedback, EF-Landing ensures convergence and constraint feasibility while significantly reducing communication overhead. We provide sharp convergence guarantees, demonstrating that EF-Landing achieves the same asymptotic linear speedup convergence rate as existing methods without communication compression. Furthermore, our analysis is highly versatile, applying to both deterministic and stochastic settings and encompassing algorithms based on gradient descent or momentum-based gradient descent. We also generalize EF-Landing to operate on block-wise Stiefel manifolds, enabling greater flexibility for structured constraints. Extensive numerical experiments validate our theoretical results. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:46:15 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Song",
"Yilong",
""
],
[
"Li",
"Peijin",
""
],
[
"Gao",
"Bin",
""
],
[
"Yuan",
"Kun",
""
]
] |
2506.02881 | Brian Cho | Brian M Cho, Aur\'elien Bibaut, Nathan Kallus | Simulation-Based Inference for Adaptive Experiments | null | null | null | null | stat.ME cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-arm bandit experimental designs are increasingly being adopted over standard randomized trials due to their potential to improve outcomes for study participants, enable faster identification of the best-performing options, and/or enhance the precision of estimating key parameters. Current approaches for inference after adaptive sampling either rely on asymptotic normality under restricted experiment designs or underpowered martingale concentration inequalities that lead to weak power in practice. To bypass these limitations, we propose a simulation-based approach for conducting hypothesis tests and constructing confidence intervals for arm specific means and their differences. Our simulation-based approach uses positively biased nuisances to generate additional trajectories of the experiment, which we call \textit{simulation with optimism}. Using these simulations, we characterize the distribution potentially non-normal sample mean test statistic to conduct inference. We provide guarantees for (i) asymptotic type I error control, (ii) convergence of our confidence intervals, and (iii) asymptotic strong consistency of our estimator over a wide variety of common bandit designs. Our empirical results show that our approach achieves the desired coverage while reducing confidence interval widths by up to 50%, with drastic improvements for arms not targeted by the design. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:46:59 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Cho",
"Brian M",
""
],
[
"Bibaut",
"Aurélien",
""
],
[
"Kallus",
"Nathan",
""
]
] |
2506.02883 | Anthony Kobanda | Anthony Kobanda, Odalric-Ambrym Maillard, R\'emy Portelas | A Continual Offline Reinforcement Learning Benchmark for Navigation Tasks | arXiv admin note: text overlap with arXiv:2412.14865 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Autonomous agents operating in domains such as robotics or video game simulations must adapt to changing tasks without forgetting about the previous ones. This process called Continual Reinforcement Learning poses non-trivial difficulties, from preventing catastrophic forgetting to ensuring the scalability of the approaches considered. Building on recent advances, we introduce a benchmark providing a suite of video-game navigation scenarios, thus filling a gap in the literature and capturing key challenges : catastrophic forgetting, task adaptation, and memory efficiency. We define a set of various tasks and datasets, evaluation protocols, and metrics to assess the performance of algorithms, including state-of-the-art baselines. Our benchmark is designed not only to foster reproducible research and to accelerate progress in continual reinforcement learning for gaming, but also to provide a reproducible framework for production pipelines -- helping practitioners to identify and to apply effective approaches. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:48:20 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Kobanda",
"Anthony",
""
],
[
"Maillard",
"Odalric-Ambrym",
""
],
[
"Portelas",
"Rémy",
""
]
] |
2506.02884 | Felipe Diaz | Jos\'e Barrientos, Nicol\'as C\'aceres, Felipe Diaz, Ulises Hernandez-Vera | ModMax Electrodynamics and Holographic Magnetotransport | 32 pages, 5 figures | null | null | null | hep-th gr-qc | http://creativecommons.org/licenses/by/4.0/ | We study magnetotransport in a holographic model where ModMax nonlinear electrodynamics is coupled to Einstein AdS gravity. To incorporate momentum relaxation, we introduce spatially linear axion fields that break translational symmetry, resulting in an anisotropic medium. Using linear response theory, we compute the DC conductivity matrix in the presence of an external magnetic field, expressing the conductivities in terms of horizon data. Our results demonstrate how the nonlinear ModMax parameter modifies charge transport, particularly influencing the Hall angle and Nernst signal. The nonlinear corrections introduce distinct deviations in both longitudinal and Hall conductivities while preserving the characteristic temperature scaling of strange metals, offering new insights into strongly coupled systems with nonlinear electromagnetic interactions. Notably, the Nernst signal reproduces that of high$-T_c$ cuprate superconductors showing a superconducting dome and a normal phase, with the ModMax deformation parameter tuning critical and onset temperatures. In the strongly nonlinear regime, we find evidence of an exotic state dominated by quasiparticle excitations in the dual material. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:50:36 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Barrientos",
"José",
""
],
[
"Cáceres",
"Nicolás",
""
],
[
"Diaz",
"Felipe",
""
],
[
"Hernandez-Vera",
"Ulises",
""
]
] |
2506.02885 | Jiangtao Shi | Jiangtao Shi, Fanjie Xu, Na Li | On an extension of Shlyk's theorem | null | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | In this paper, we prove that the intersection of all non-nilpotent maximal subgroups of a non-solvable group containing the normalizer of some Sylow subgroup is nilpotent, which provides an extension of Shlyk's theorem. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:51:03 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Shi",
"Jiangtao",
""
],
[
"Xu",
"Fanjie",
""
],
[
"Li",
"Na",
""
]
] |
2506.02886 | Sakshi Gupta | Sakshi Gupta, Souvik Roy and Dibyendu Nandy | Stellar Magnetic Storm Induced Magnetospheric Polarity Reversals: Distinguishing between Unmagnetised and Magnetised Exoplanets | Submitted | null | null | null | astro-ph.EP astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Exoplanetary and planetary environments are forced by stellar activity which manifest through variable radiation, particle and magnetic fluxes, stellar winds, flares and magnetic storms known as coronal mass ejections (CMEs). Recent studies have shown that (exo)planets with intrinsic magnetic fields and magnetospheres respond differently to this stellar forcing compared to planets which lack an intrinsic magnetism; this is borne out by observations in solar system planets. However, detailed investigations to uncover the subtle ways in which stellar magnetic storms impact exoplanets are still at a nascent stage. Here we utilize 3D magnetohydrodynamic simulations to investigate the impact of stellar CMEs on Earth-like planets with different magnetic fields. Our results show that planetary atmospheric mass loss rates are dependent on the relative orientation of stellar wind and planetary magnetic fields, with significantly higher losses when the CME and planetary magnetic fields are oppositely oriented -- favoring enhanced magnetic reconnections. In contrast, for unmagnetised planets, the mass loss rate do not strongly depend on stellar magnetic field orientation. More significantly, we find that stellar CME induced polarity reversals can distinguish between planets with and without intrinsic magnetism. In unmagnetised or weakly magnetised (exo)planets, the polarity of the externally imposed magnetosphere are prone to global polarity reversals forced by stellar magnetised storms. Our analysis of the magnetotail current density dynamics during polarity reversals aligns with observations of Venus. This distinction in magnetospheric response provides a new paradigm to differentiate between (exo)planets with or without significant (intrinsic) magnetic fields. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:52:19 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Gupta",
"Sakshi",
""
],
[
"Roy",
"Souvik",
""
],
[
"Nandy",
"Dibyendu",
""
]
] |
2506.02887 | Rohit Agarwal | Mrinmay Sen, Shruti Aparna, Rohit Agarwal, Chalavadi Krishna Mohan | Overcoming Challenges of Partial Client Participation in Federated Learning : A Comprehensive Review | 15 pages, 6 tables, comprehensive survey of federated learning with partial client participation | null | null | null | cs.LG cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated Learning (FL) is a learning mechanism that falls under the distributed training umbrella, which collaboratively trains a shared global model without disclosing the raw data from different clients. This paper presents an extensive survey on the impact of partial client participation in federated learning. While much of the existing research focuses on addressing issues such as generalization, robustness, and fairness caused by data heterogeneity under the assumption of full client participation, limited attention has been given to the practical and theoretical challenges arising from partial client participation, which is common in real-world scenarios. This survey provides an in-depth review of existing FL methods designed to cope with partial client participation. We offer a comprehensive analysis supported by theoretical insights and empirical findings, along with a structured categorization of these methods, highlighting their respective advantages and disadvantages. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:52:27 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Sen",
"Mrinmay",
""
],
[
"Aparna",
"Shruti",
""
],
[
"Agarwal",
"Rohit",
""
],
[
"Mohan",
"Chalavadi Krishna",
""
]
] |
2506.02888 | Ignacio Gonz\'alez | Ignacio Gonz\'alez, Sagnik Chakraborty, \'Angel Rivas | Intrinsic Hamiltonian of Mean Force and Strong-Coupling Quantum Thermodynamics | RevTex4 File, Color Figures | null | null | null | quant-ph cond-mat.mes-hall cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a universal thermodynamic framework for quantum systems that may be strongly coupled to thermal environments. Unlike previous approaches, our method enables a clear definition of thermostatic properties while preserving the same gauge freedoms as in the standard weak-coupling regime and retaining the von Neumann expression for thermodynamic entropy. Furthermore, it provides a formulation of general first and second laws using only variables accessible through microscopic control of the system, thereby enhancing experimental feasibility. We validate the framework by applying it to a paradigmatic model of strong coupling with a structured bosonic reservoir. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:54:07 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"González",
"Ignacio",
""
],
[
"Chakraborty",
"Sagnik",
""
],
[
"Rivas",
"Ángel",
""
]
] |
2506.02889 | Xi Wang | Ella Xi Wang, Melissa Ness, Thomas Nordlander, Andrew R. Casey, Sarah Martell, Marc Pinsonneault, Xiaoting Fu, Dennis Stello, Claudia Reyes, Marc Hon, Madeleine McKenzie, Mingjie Jian, Jie Yu, Sven Buder, Karin Lind, Joss Bland-Hawthorn, Daniel B. Zucker, Pradosh Barun Das, Richard de Grijs, Michael Hayden | Optical spectroscopic signatures of the red giant evolutionary state | 16 pages, 13 figures, accepted for publication in MNRAS | null | null | null | astro-ph.SR astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern spectroscopic surveys output large data volumes. Theoretical models provide a means to transform the information encoded in these data to measurements of physical stellar properties. However, in detail the models are incomplete and simplified, and prohibit interpretation of the fine details in spectra. Instead, the available data provide an opportunity to use data-driven, differential analysis techniques, as a means towards understanding spectral signatures. We deploy such an analysis to examine core helium-fusing red clump (RC) and shell hydrogen-fusing red giant branch (RGB) stars, to uncover signatures of evolutionary state imprinted in optical stellar spectra. We exploit 786 pairs of RC and RGB stars from the GALAH survey, chosen to minimise spectral differences, with evolutionary state classifications from TESS and K2 asteroseismology. We report sub-percent residual, systematic spectral differences between the two classes of stars, and show that these residuals are significant compared to a reference sample of RC$-$RC and RGB$-$RGB pairs selected using the same criteria. First, we report systematic differences in the Swan ($\rm{C}_2$) band and CN bands caused by stellar evolution and a difference in mass, where RGB stars at similar stellar parameters have higher masses than RC stars. Secondly, we observe systematic differences in the line-width of the H$_{\alpha}$ and H$_{\beta}$ lines caused by a difference in microturbulence, as measured by GALAH, where we measure higher microturbulence in RC stars than RGB stars. This work demonstrates the ability of large surveys to uncover the subtle spectroscopic signatures of stellar evolution using model-free, data-driven methods. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:54:49 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Wang",
"Ella Xi",
""
],
[
"Ness",
"Melissa",
""
],
[
"Nordlander",
"Thomas",
""
],
[
"Casey",
"Andrew R.",
""
],
[
"Martell",
"Sarah",
""
],
[
"Pinsonneault",
"Marc",
""
],
[
"Fu",
"Xiaoting",
""
],
[
"Stello",
"Dennis",
""
],
[
"Reyes",
"Claudia",
""
],
[
"Hon",
"Marc",
""
],
[
"McKenzie",
"Madeleine",
""
],
[
"Jian",
"Mingjie",
""
],
[
"Yu",
"Jie",
""
],
[
"Buder",
"Sven",
""
],
[
"Lind",
"Karin",
""
],
[
"Bland-Hawthorn",
"Joss",
""
],
[
"Zucker",
"Daniel B.",
""
],
[
"Das",
"Pradosh Barun",
""
],
[
"de Grijs",
"Richard",
""
],
[
"Hayden",
"Michael",
""
]
] |
2506.02890 | Jakub Krajewski | Jakub Krajewski, Marcin Chochowski, Daniel Korzekwa | Scaling Fine-Grained MoE Beyond 50B Parameters: Empirical Evaluation and Practical Insights | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Mixture of Experts (MoE) architectures have emerged as pivotal for scaling Large Language Models (LLMs) efficiently. Fine-grained MoE approaches - utilizing more numerous, smaller experts - have demonstrated potential in improving model convergence and quality. This work proposes a set of training recipes and provides a comprehensive empirical evaluation of fine-grained MoE, directly comparing its scaling properties against standard MoE configurations for models with up to 56B total (17B active) parameters. We investigate convergence speed, model performance on downstream benchmarks, and practical training considerations across various setups. Overall, at the largest scale we show that fine-grained MoE achieves better validation loss and higher accuracy across a set of downstream benchmarks. This study offers empirical grounding and practical insights for leveraging fine-grained MoE in the development of future large-scale models. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:55:48 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Krajewski",
"Jakub",
""
],
[
"Chochowski",
"Marcin",
""
],
[
"Korzekwa",
"Daniel",
""
]
] |
2506.02891 | Leena Mathur | Jiewen Hu, Leena Mathur, Paul Pu Liang, Louis-Philippe Morency | OpenFace 3.0: A Lightweight Multitask System for Comprehensive Facial Behavior Analysis | IEEE FG 2025, \c{opyright} 2025 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, there has been increasing interest in automatic facial behavior analysis systems from computing communities such as vision, multimodal interaction, robotics, and affective computing. Building upon the widespread utility of prior open-source facial analysis systems, we introduce OpenFace 3.0, an open-source toolkit capable of facial landmark detection, facial action unit detection, eye-gaze estimation, and facial emotion recognition. OpenFace 3.0 contributes a lightweight unified model for facial analysis, trained with a multi-task architecture across diverse populations, head poses, lighting conditions, video resolutions, and facial analysis tasks. By leveraging the benefits of parameter sharing through a unified model and training paradigm, OpenFace 3.0 exhibits improvements in prediction performance, inference speed, and memory efficiency over similar toolkits and rivals state-of-the-art models. OpenFace 3.0 can be installed and run with a single line of code and operate in real-time without specialized hardware. OpenFace 3.0 code for training models and running the system is freely available for research purposes and supports contributions from the community. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:56:10 GMT"
}
] | 2025-06-04T00:00:00 | [
[
"Hu",
"Jiewen",
""
],
[
"Mathur",
"Leena",
""
],
[
"Liang",
"Paul Pu",
""
],
[
"Morency",
"Louis-Philippe",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.