The paper introduces Skeletal Prototypes on Iterative Nerve Expansions (SPINE), a prototype reduction method that represents each class as an embedded 1‑complex rather than a finite set of points. SPINE constructs its initial edge set from a class‑conditional Mapper graph, then refines vertex positions under a classification objective, allowing observations to be assigned to the nearest complex. Evaluated on seventeen benchmark datasets with stratified 10‑fold cross‑validation, SPINE achieves the highest mean accuracy and best average rank among seven competing methods, showing significant improvements over five of them and competitive performance across varying prototype budgets.
By Jordan Eckert, Henry Schenck
The paper introduces a new parameter pruning scheme based on the Fisher information metric, using the geodesic distance in model space to determine the minimal change when a parameter is set to zero. This approach creates a hierarchy of pruning methods, starting from traditional magnitude pruning and progressing to more sophisticated, effective schemes. Experiments on fully‑connected networks and vision transformers across MNIST and CIFAR‑10 show that the Fisher‑based method outperforms magnitude pruning and local Fisher pruning in accuracy and Matthews correlation coefficient, while intermediate approximations offer computational efficiency with near‑optimal performance.
By David S. Berman, Yen-Yu Fu, Edward Hirst, Thelma Chiwete Obirai
The paper introduces Affect-Prototype Guided Fusion (APCF), a framework for open‑vocabulary multimodal emotion recognition that handles incomplete and unsynchronized modal data. APCF builds an affect‑prototype library to model how different emotions contribute across modalities, enabling dynamic fusion of available features. The fused representations are then decoded by an LLM to generate open‑vocabulary emotion labels, achieving superior performance on OV‑MERD+ and MER‑FG datasets compared to existing methods.
By Yichi Zhang, Shenyue Wang, Jing Luo, Chunyang Yu, Xinyu Yang
The paper introduces Symbolic Separation, a method that grounds deep learning agents in knowledge graphs to improve reliability in operational data analytics. By restricting agent actions to an ontology-constrained Virtual Knowledge Graph with deterministic pre-execution validation, the approach transforms complex queries into validated graph traversals rather than relying on LLM-inferred joins. In experiments on 49.9 TB of supercomputer telemetry, the Neurosymbolic Deep Analyst achieved an 86% task‑success rate, eliminated silent data‑integrity errors, and reduced token costs by 2.4× compared to a non‑symbolic baseline.
By Baibek Davletiyarov, Junaid Ahmed Khan, Andrea Bartolini
EviDep is a multimodal evidential regression framework for estimating depression severity from audio–visual recordings, incorporating multi‑scale temporal modeling and shared–private representation learning. It uses frequency‑aware feature extraction to decompose behavioral sequences into multiple frequency bands, refined by scale‑specific experts, and applies disentangled evidential learning to separate cross‑modal shared and modality‑specific information. The model outputs Normal‑Inverse‑Gamma distributions via multi‑branch evidential regression, enabling estimation of depression severity along with aleatoric and epistemic uncertainty, and demonstrates competitive accuracy on several benchmark datasets.
By Fangyuan Liu, Sirui Zhao, Yangsong Zhang, Jinyang Huang, Feng-Qi Cui, Bin Luo, Tong Xu, Enhong Chen
The paper presents a method for compressing large tabular foundation models into compact feed‑forward students for use in interactive agentic decision loops. By distilling a TabPFN teacher into a student, the authors reduce model size from 53.2 M to 8,546 parameters (6,220× compression) and from 111.4 M to 17,059 parameters (6,532× compression) while preserving 95.4–100.5% accuracy and 96.8–100.0% AUC across several benchmarks. The study demonstrates that soft‑target distillation yields a 2.1–7.0 AUC improvement over hard‑label training.
By Sourish Dey, Aditya Kumar
arXiv:2507.23248v2 Announce Type: replace-cross
Abstract: Bengali is spoken by more than 230 million people, yet no standardized instrument evaluates large language models (LLMs) on Bengali across th...
By Shimanto Bhowmik, Tawsif Tashwar Dipto, Md Sazzad Islam, Sheryl Hsu, Tahsin Reasat
arXiv:2510.09619v2 Announce Type: replace-cross
Abstract: [Corrected v2: an audit found that the score, threshold, and latency descriptions below are not what the shared codebase implements, and that...
By Michel A. Youssef (Independent Researcher)
arXiv:2609.16739v1 Announce Type: new
Abstract: Background: Large language models (LLMs) have achieved physician-comparable performance on multiple-choice medical knowledge examinations, but their ca...
By Keisuke Masuda, Kazutaka Yatsushiro, Hirohumi Iwamoto, Hirofumi Hirano, Ryosuke Hanaya
arXiv:2609.16006v1 Announce Type: cross
Abstract: Large language models (LLMs) increasingly serve users whose expectations are shaped by their cultural context, yet most cultural evaluations test wha...
By Enes Altinisik, Hamdy Mubarak, Masoomali Fatehkia, Husrev_Taha_Sencar Husrev Taha Sencar
The paper introduces little m, an AI agent that helps formulate industrial process control models by combining a domain-specific knowledge repository with LLM-driven interaction. It tackles the challenge of converting messy real-world specifications, including natural language and spatial diagrams, into rigorous mathematical optimization models. The authors also present IPC-Bench, a multimodal dataset of 50 canonical scenarios, and show through automated and human evaluations that little m outperforms state‑of‑the‑art LLMs in generating semantically correct models.
By Yongchao Ye, Xinyu He, Dutliff Boshoff, Way Kuo, Lishuai Li
The paper introduces MoE-JEPA, a dual‑stream deepfake detection model that combines a V‑JEPA backbone with a Residual Mixture‑of‑Experts mechanism and a noise stream branch. It further incorporates a Gated Attention Multiple Instance Learning module to refine spatial semantic understanding. On the SID‑Set benchmark, MoE‑JEPA achieves a new state‑of‑the‑art accuracy of 95.54%, outperforming much larger models.
By Simone Teglia, Irene Amerini
High-Fidelity Video Quality Assessment (HFVQA) is a new framework that uses fixed-size spatio‑temporal patches across multiple scales, including the original resolution, to preserve low‑level quality cues and semantic context. It incorporates a lightweight auxiliary network that learns VQA‑specific saliency directly from quality supervision, enabling the model to focus on the most important spatio‑temporal regions. By combining high‑fidelity cues with task‑specific saliency, HFVQA achieves state‑of‑the‑art performance on standard no‑reference VQA benchmarks while processing only about 12% of the candidate patches, making it computationally efficient.
By Hakan Emre Gedik, Shashank Gupta, Alan Bovik
arXiv:2609.16800v1 Announce Type: new
Abstract: Large Language Models (LLMs) have achieved remarkable progress across diverse domains, but continual adaptation to evolving tasks and environments rema...
By Ting-Wei Chang, Po-Chun Chen, Hen-Hsen Huang, Hsin-Hsi Chen
The paper introduces the first systematic robustness benchmark for local invisible image watermarking, evaluating five methods across 55 image transformations that include signal distortions, coordinate misalignments, indirect local edits, and direct watermark edits. Results show that all methods are vulnerable to some transformation, with MaskWM achieving the best payload recovery and localization but at the cost of image quality. The study highlights that robustness varies strongly with transformation type, especially noting that geometric misalignment and generative local edits can completely disrupt payload recovery.
By Kai Yao, Bence Szil\'agyi, Sebesty\'en Kamp, M\'at\'e Po\'or, M\'at\'e Szilveszter, Matyas K. Zsoldos, Marc Juarez
ECHO is a new paired diagnostic benchmark designed for Chinese full‑duplex spoken dialogue systems to evaluate context‑sensitive turn‑taking. It pairs examples that share the same overlap transcript but differ in preceding multi‑turn context, requiring either a Yield or Keep decision, and also includes off‑talk cases to test unnecessary yielding. The benchmark introduces pair accuracy, penalizing constant‑action policies, and shows that many systems are biased toward Yield, performing better on interruptions than on backchannels.
By Shuofeng Zhao, Hongwei Cai, Wenke Fan, Qingxiang Guo, Dawei Yang, Zhou Wang, Zhiyang Zhou, Yingxin Shang, Weixu Wang, Lin Yang, Shuran Zhou, Yang Song
arXiv:2609. 17435v1 Announce Type: new Abstract: We submit M\'eTRON-FR, a 125M GPT-2 pretrained on 92.
By Adam Zachary Wasserman, David Beauchemin
SciNLP is a new benchmark dataset for full‑text entity and relation extraction in the NLP domain, comprising 60 manually annotated papers with 6,429 entities and 1,649 relations. It is the first dataset to provide full‑text annotations of entities and their relationships specifically for NLP literature. Experiments show that models trained on SciNLP outperform baselines on certain tasks, and the dataset enabled the automatic construction of a fine‑grained knowledge graph with an average node degree of 3.3.
By Decheng Duan, Yingyi Zhang, Jitong Peng, Chengzhi Zhang
SOTER is a generative foundation model designed for wearable physiological time‑series data. It integrates cross‑channel coupling, spectrum‑guided expert specialization, and continuous‑time latent evolution, using a spatial feature‑aware backbone, a PSD‑guided mixture‑of‑experts layer, and a neural controlled differential equation decoder. Trained on 226 billion time points from five public datasets, SOTER outperforms baselines in zero‑shot forecasting, classification, and imputation across six benchmarks, and remains robust to additive noise.
By Fangke Chen, Sirry Chen, Wei Chen, Zhongyu Wei
EgoPathBench is a new dataset and benchmark that tests zero‑shot egocentric waypoint decision‑making in vision‑language models. Each task presents an egocentric RGB image, a natural‑language goal, and numbered visible waypoints, and models must return traversable candidates or an ordered route. The benchmark evaluates candidate feasibility, edge legality, and goal arrival under point‑agent or embodied geometry, covering 31,852 training, 1,345 validation, and 1,111 benchmark questions.
"whyItMatters":"The benchmark reveals that current VLMs perform poorly on integrated navigation tasks, highlighting a gap in spatial intelligence that can be addressed by fine‑tuning with the released training data."
By Yang Zhao, Zhuo Chen, Xubo Yang