The paper introduces Message‑Passing Graph Persistence (MGP), a new family of topological descriptors that integrates message‑passing into graph filtrations. By updating node embeddings at each filtration step and using them to determine persistence events and component representatives, MGP remains invariant under graph isomorphism and is proven to be strictly more expressive than existing PH‑based descriptors. Experiments on expressivity benchmarks and various graph classification and regression datasets demonstrate that MGP outperforms other PH‑based methods on most tasks.
By Jorge Luiz Franco, Arthur Queiroz Moura, Moacir Ponti, Amauri Souza
Bidirectional Centered Score Matching (BCSM) is introduced as an efficient within-position distillation objective for discrete diffusion models, reducing the complexity from ≠|V|^2 to ≠K^2 by using an ≠K bidirectional active set. BCSM aligns relative logit geometry while maintaining a factorized ≠K objective, and it incorporates cross-timestep transition alignment. Experiments show significant efficiency gains, achieving up to 6× fewer neural function evaluations and notable improvements in HumanEval and MMMU metrics at single-step inference.
By Peng Tu, Baoyou Chen, Haoyuan Xia, Zilong Dong, Weihao Yuan, Siyu Zhu
MESA is an evaluation framework that precisely measures language‑conditioned policy generalization in robotics. It offers five evaluation suites that separate in‑distribution performance from four axes of generalization—spatial configuration, object instance, object category, and subtask composition—allowing researchers to pinpoint why policies fail. The authors also introduce MESA‑Gen, a task and demonstration generation pipeline, and release two benchmark suites: a single‑arm MESA and BiMESA for bimanual generalization, with baseline results showing current models achieve only about 50% success, indicating significant room for improvement.
By Albert Wilcox, Frank Chang, Nhi Nguyen, Aishani Chakraborty, Jeremy A. Collins, Vaibhav Saxena, Benjamin Joffe, Siddharth Karamcheti, Animesh Garg
The paper introduces a Multiscale Entanglement Renormalization Ansatz (MERA)-inspired autoencoder for anomaly detection in collider jets, leveraging the jets’ natural multiscale structure. It compares this architecture to dense autoencoders, tree-tensor-network limits, and classical baselines, finding that the MERA model achieves higher mean AUC with fewer trainable parameters. The study demonstrates that locality-aware hierarchical compression and disentangling layers provide a useful, parameter‑efficient inductive bias for reconstruction‑based jet anomaly detection.
By Emre Gurkanli, Michael Spannowsky
The paper introduces “LExMem”, a graph‑guided benchmark that evaluates long‑term memory in exploration tasks by modeling these processes as evolving graphs and generating 224 dialogues with 2,018 question‑answer pairs. It also presents “LExGraph”, a graph‑based memory framework that incrementally organizes dialogues into interconnected nodes and retrieves evidence via hybrid retrieval and neighbor expansion. Experiments show that “LExMem” challenges existing systems, while “LExGraph” outperforms them and generalizes to external benchmarks.
By Chongyuan Dai, Yaling Shen, Zirui Song, Junfeng Liao, Jinpeng Hu, Xiuying Chen
The paper introduces a visual-native agent harness that registers every tool‑returned image as an addressable reference, enabling reuse of intermediate visual evidence in multimodal deep search tasks. It also proposes On‑Policy Data Evolution (ODE), a closed‑loop data generator that refines training data across rounds based on policy rollouts, targeting the agent’s evolving needs. Experiments on eight benchmarks show ODE boosts the Qwen3‑VL‑8B agent’s performance from 24.9% to 39.0% on average, and from 30.6% to 41.5% at 30B, while validating the benefits of image bank reuse and rollout‑feedback evolution.
By Shijue Huang, Hangyu Guo, Guanting Dong, Chenxin Li, Junting Lu, Xinyu Geng, Zhaochen Su, Zhenyu Li, Shuang Chen, Hongru Wang, Yi R. Fung
GeoBiaset is a new benchmark that tests how world-level image geolocalization models react to the apparent race of a person inserted into the image foreground. The dataset contains 9,541 images, each paired with counterfactual edits across seven races while keeping the background and ground-truth coordinates unchanged. Experiments show that demographic perturbations can cause a 16–47% drop in continent-level accuracy, and that models shift their focus from geographic landmarks to facial features when landmarks are absent.
By Niccol\`o Niccoli, Silvia Dani, Marco Mistretta, Andrew D. Bagdanov, Marco Bertini, Lorenzo Seidenari
SSFMamba is a Mamba-based framework that fuses spatial and frequency-domain modeling for 3D medical image segmentation. It uses a dual-branch design, with a spatial branch preserving voxel-level textures and a frequency branch capturing global context through spectral components. The core 3D Multi-Directional Scanning Mechanism incorporates Hermitian symmetry and State Space Models to enable direction-aware global modeling, leading to superior performance on MRI and CT datasets, especially for low-contrast organs like the pancreas.
By Bo Zhang, Yifan Zhang, Shuo Yan, Yu Bai, Zheng Zhang, Wu Liu, Xiuzhuang Zhou, Wendong Wang
SlackBench is a new benchmark that evaluates AI agents on realistic collaborative research and engineering workflows, requiring joint reasoning over workplace communication and evolving repository state. The benchmark is built from privacy‑preserving reenactments of real academic projects, providing reproducible environments with multi‑party dialogue, code changes from GitHub, and Google Docs. It contains 140 queries covering heterogeneous‑source search, abstention, summarization, and inference over long conversations, and shows that current models perform well below the ceiling, indicating significant room for improvement.
By Mingxuan Li, Huanzhi Mao, Ruitao Zheng, Jarod Stanbury James, Yunze Li, Natanya Anderson, Joseph E. Gonzalez, Alex Dimakis
The paper introduces SafeStart, a technique for Mixture-of-Experts (MoE) language models that mitigates the safety degradation caused by active expert reduction during serving. By running all experts for the prompt and the first eight generated tokens before switching to a reduced set, SafeStart largely restores the model’s refusal to harmful requests while preserving most of the speedup achieved by expert reduction. Experiments on several MoE models show that SafeStart maintains high offline decode speedups (≈85% of the original) and limits the increase in harmful responses to within a few percentage points, even under jailbreak prompts.
By Erfan Baghaei Potraghloo, Saeid Shokoufa, Seyedarmin Azizi, Arya Fayyazi, Souvik Kundu, Mehdi Kamal, Massoud Pedram
The paper introduces Coca-Cola, a continuous latent context compression model that transforms raw tool outputs into latent tokens for LLM agents, bypassing the need for explicit text representation. By embedding these latent tokens into the agent’s input space and employing Action‑Preserving Training, the method ensures that decision‑critical information is retained while reducing token consumption by 70.01%–89.24%. Experiments on web‑search agents across multiple QA benchmarks demonstrate significant reductions in runtime (24.45%–55.97%) and consistent improvements in task performance.
By Yuetong Liu, Zhehao Zhou, Yitong Zhang, Haorui Chen, Jia Li
RAGAM-AI is a hybrid framework that merges reinforcement learning (RL) with Generative AI to manage resource allocation in Multi-access Edge Computing (MEC) environments. It addresses reward generalization challenges by employing a Large Language Model (LLM) for feedback and uses a Multi-Agent Soft-Actor Critic (MASAC) with a surrogate reward model to guide resource orchestration across distributed edge servers. Compared to single-agent SAC, REINFORCE, and random policies, RAGAM-AI achieves lower deviations in CPU, memory, and bandwidth allocation while reducing latency and data rate costs.
By Shreya K. Chari, John S. Vardakas, Kostas Ramantas, Christos Verikoukis
The paper introduces MechScam, a benchmark of 6,614 videos that covers diverse scam mechanisms and non‑scam content, providing evidence‑supported explanations and annotations of scam elements. It also proposes ScamCourt, a debate framework where prosecution and defense debaters argue over verified citations, and a judge decides based on the transcript. Experiments show that current large multimodal language models struggle to differentiate scam videos from similar non‑scam content, while ScamCourt reduces false positives and better resists misleading arguments.
By Yifan Wang, Junyu Lu, Shun Zhang, Chaozhuo Li, Fanliang Bu, Rong Yin, Lingbin Bu, Changrong Min, Peng Chen
CAST-Bench is a new benchmark for multi-shot audio‑video generation that traces each generated sound back to its intended source. It contains 900 prompts across nine sound‑source categories, including both single‑shot and multi‑shot narratives, and evaluates models using 18 metrics grouped into Visual Performance, Content Consistency, and Audio & Semantic Alignment. The benchmark reveals that while visual quality is largely saturated, audio semantic scores vary significantly, and multi‑shot prompts expose challenges such as consistent speaker identity and cross‑shot event retention.
By Yucong Liu, Qi Tian, Shuyuan Tu, Kaihang Pan, Weijie Kong, Chao Pan
The paper introduces ZETA, an architecture that separates routing policy from action execution in scientific vision agents. By resolving scientific intent before invoking specialized visual tools, ZETA treats measurement as a deterministic operation grounded in explicit geometry, enabling independent evaluation of routing correctness and downstream execution. Experiments on a zero‑shot benchmark of 51 scientific‑intent queries demonstrate that explicit routing semantics can dramatically improve routing accuracy, especially for models like Qwen 3.8 where segmentation routing accuracy jumps from 6.2% to 75.0% and detection routing accuracy from 60.0% to 100%.
By Shubhabrata Mukherjee, Jeffrey Donatelli, Daniela Ushizima
arXiv:2507.22186v3 Announce Type: replace-cross
Abstract: Given a query table, existing data lake discovery techniques seek to identify data sources related to the query table through schema similari...
By Ambarish Singh, Romila Pradhan
arXiv:2508.06199v5 Announce Type: replace-cross
Abstract: Pretrained molecular embedding models spanning graph neural networks, text transformers, and hybrid architectures have been widely used in dr...
By Mateusz Praski, Jakub Adamczyk, Wojciech Czech
arXiv:2602.09220v2 Announce Type: replace-cross
Abstract: Time series forecasting is a critical task across domains such as energy, finance, and meteorology, where accurate predictions enable informe...
By Julien Guit\'e-Vinet, Alexandre Blondin Mass\'e, \'Eric Beaudry
arXiv:2602.20207v4 Announce Type: replace-cross
Abstract: Knowledge editing in Large Language Models (LLMs) aims to update the model's prediction for a specific query to a desired target while preser...
By Shrestha Datta, Hongfu Liu, Anshuman Chhabra
arXiv:2603.13275v4 Announce Type: replace-cross
Abstract: Accurate prediction of surgical duration is pivotal for hospital resource management. Although recent supervised learning approaches, from ma...
By Wanyin Wu, Kanxue Li, Baosheng Yu, Haoyun Zhao, Yibing Zhan, Dapeng Tao, Hua Jin