CuratorMAS is a multi‑agent framework that automates dataset curation by orchestrating agents through five programmable stages: exploration, knowledge retrieval, criterion derivation, filtering, and evolution. It gathers contextual information, augments evaluation with online domain knowledge, computes metrics, filters data, and updates skills based on outcomes. Experiments show it can lower noise rates by up to 36.03 pp and boost downstream model F1 scores by up to 8.88 pp.
By Yixin Zhang, Wenjie Feng
The paper "Auditable Claims about AI Agents" argues that for a claim about an AI agent to be verifiable, it must explicitly state the agent’s policy, scope, the records that would confirm the claim, and who is responsible for those records. It extends the Auditable Agents framework by adding conditions of independent record coverage, action‑specific authorization, and completeness beyond integrity, and demonstrates the approach with a claim‑check table applied to six common claims. The authors provide a worked example and practical guidance for operators, buyers, auditors, and standard setters.
By Yue Zhao, Jiate Li, Li Li, Yi Nian, Jinbo Liu, Xiaolin Zhou, Xiyang Hu
The paper introduces HARP, a hard-budget allocation method for evaluating large language models (LLMs) in multi-turn interactions where the time-to-event is partially observed due to resource limits. HARP guarantees that the total computational budget is never exceeded, reallocates unused budget, and provides lower predictive bounds (LPBs) with finite-sample coverage and unbiased metric estimates. Experiments on tasks such as jailbreaks, toxic content, and hallucinations demonstrate that HARP achieves near-nominal coverage with low variance while respecting the fixed budget.
By Shai Feldman, Yaniv Romano
The paper investigates whether large language models (LLMs) can infer the capability requirements of a user query before generating a response. Using the TACIT framework, the authors decompose these requirements into eight classes across three axes—Source, Transformation, and World Effect—and train linear probes on hidden states from four open-weight LLM families. Their results show that these capability structures are linearly decodable with high accuracy, yet the models struggle to express the same information in natural language, a phenomenon termed "structured but silent."
By Kyojun Choo, Minsoo Song, Yunju Kang, Chanjun Park
The paper proposes a production‑centered framework for comparing detection and watermarking techniques across two main AI image and video creation routes: direct visual generation and LLM‑driven code rendering. It introduces an explicit verification specification that separates passive inference, message recovery, and authenticated provenance, and organizes watermarks by production stage for images, videos, source code, and rendering‑aware outputs. The authors outline ten research questions covering identifiability, observability, fair comparison, payload recoverability, reconstruction, synchronization, composition, hybrid local contribution, and private production‑event authentication, and connect the framework to concrete systems such as Claude, OpenAI, and rendering tools.
By Zheng Gao, Xiaoyu Li, Zhicheng Bao, Yang Song, Jiaojiao Jiang
The paper introduces FindIt, the first comprehensive benchmark for evaluating the promptable localization abilities of generalist multimodal large language models (MLLMs). It covers four core task categories—object detection, referring expression detection, instance-level detection, and video-based detection—and provides a unified framework that standardizes inputs, enforces parsable bounding box outputs, and defines transparent evaluation protocols. Using this benchmark, the authors assess a range of open-source and proprietary MLLMs, revealing both their strengths and limitations, particularly their sensitivity to formatting constraints and difficulty generalizing to minor variations.
By Eshika Khandelwal, Jingjing Pan, Mingfang Zhang, Quan Kong, Lorenzo Garattoni, Hilde Kuehne
arXiv:2610.07132v1 Announce Type: cross
Abstract: Croissant has emerged as a standard for machine-readable dataset metadata, yet populating its fields remains labor-intensive and requires careful rea...
By Berke Arda, Ahmetcan Yavuz, Paul Gerry, Sebastian Lobentanzer, Nobin Sarwar, Joan Giner-Miguelez, Kongtao Chen, Luyao Zhang, Mrinmaya Sachan, Mubashara Akhtar
arXiv:2610.05281v2 Announce Type: replace
Abstract: Modern agents increasingly ground their reasoning in observations returned by tools, such as file contents read from a workspace. However, the data...
By Yingying Liu, Junzhou Fang, Chenxiong Qian
arXiv:2610.06890v1 Announce Type: new
Abstract: We present an industry experience report on three years of operating an event-driven cloud infrastructure for continuous machine learning training in a...
By Zhengyang (Cissy), Gu, Thomas Cook, Fredaljohn Rohrbaugh, Joseph E. Hernandez, Chris Couch
arXiv:2610.07447v1 Announce Type: new
Abstract: Autoresearch agents tackle open-ended problems by repeatedly proposing candidate solutions, evaluating them, and using feedback to guide subsequent exp...
By Ziyang Cai, Christos Ziakas, Vasilis Kontonis, Tim Pearce, Siddhartha Sen, Akshay Krishnamurthy, Shivam Garg, Dimitris Papailiopoulos
arXiv:2610.07967v1 Announce Type: new
Abstract: As large language model (LLM) agents become increasingly autonomous, they may pursue task performance through deception, raising concerns about their r...
By Yiming Xu, Hongyue Yu, Beihua Yang, Zihan Chen, Yixin Liu, Zhen Peng, Bin Shi, Bo Dong, Chao Shen, Irwin King, Qinghua Zheng
arXiv:2610.08271v1 Announce Type: new
Abstract: Classical reinforcement learning (RL) assumes that a reward is observed for every visited state-action pair. However, in real-world applications such a...
By Fengxu Liu, Siwei Wang, Gal Dalal, Shie Mannor, Yihan Du
arXiv:2610.07167v1 Announce Type: cross
Abstract: We propose an imitation-learning design for neural-network control policies under state and input constraints. Training alternates a standard imitati...
By Shengfan Cao, Francesco Borrelli
arXiv:2509.18169v4 Announce Type: replace
Abstract: Tasks on complex systems require high-precision numerical computation to support decisions. However, current large language models (LLMs), even wit...
By Jingyuan Fan, Purui Liu, Hengbo Xiao, Yuxuan Zheng, Jingzhao Zhang, Chao Lu, Guannan He
arXiv:2609.34060v2 Announce Type: replace
Abstract: Prompt-specialized multi-agent systems enable multiple agents to share a model while performing complementary roles to solve complex tasks. However...
By Hyesung Jeon, Hyeongju Ha, Seoyoung Lee, Beomseok Kang, Jae-Joon Kim
arXiv:2609.34457v2 Announce Type: replace
Abstract: Transformer-based models are widely used for reasoning, coding, and multimodal agentic tasks. To provide formal assurance of desirable behaviors, s...
By Hai Duong, Thanh Le, ThanhVu Nguyen
The paper introduces CASTLE, a framework for fully decentralized multi‑agent reinforcement learning that uses two offline‑trained world models to provide in‑context guidance to independent agents. A Local Dynamics World Model captures each agent’s trajectory dynamics, while a Semantic‑Social World Model predicts short‑horizon task and social outcomes for candidate actions based on counterfactual simulator rollouts. During online learning, these frozen models supply prediction logits to an independent PPO policy, enabling agents to compare potential actions prospectively rather than relying solely on scalar rewards. CASTLE outperforms existing baselines on several multi‑particle benchmark tasks, achieving the highest mean final scores across 30 seeds.
By Fernando Martinez, Tao Li, Yingdong Lu, Juntao Chen
arXiv:2610.04687v2 Announce Type: replace
Abstract: Answering questions over imperfect tables requires handling errors that can affect the answer. We investigate two challenges for large language mod...
By Baowen Zhang, Wei Fan, Ruman Wang, Hangting Ye
arXiv:2610.07031v1 Announce Type: new
Abstract: Recent video world models have witnessed the paradigm shift from single-agent to multi-agent involvements, which can reveal more complicated dynamics a...
By Sitian Shen, Jiuming Liu, Mengmeng Liu, Yian Wang, Michael Ying Yang, Francesco Nex, Hao Cheng, Daniele De Martini, Ayush Tewari, Per Ola Kristensson
arXiv:2610.07117v1 Announce Type: new
Abstract: Understanding and modeling interactions between autonomous agents is a fundamental challenge in robotics, with broad implications for collaborative sys...
By Gabriele Magrini, Riccardo Catalini, Federico Becattini, Guido Borghi, Pietro Pala, Roberto Vezzani, Lorenzo Seidenari