The paper investigates whether response safety can be measured by the cosine similarity between a response embedding and the mean embedding of known‑safe responses. Using four frozen encoders and prompt‑controlled datasets, the authors find that a simple prototype (mean safe embedding) performs poorly (ROC‑AUC 0.457‑0.545) while an explicit safe‑minus‑unsafe reference achieves higher scores (0.588‑0.738). The study shows that a class mean is merely a location, not a safety direction, and that a reference with sufficient unsafe mass is needed to orient safety judgments.
By Sahil Kadadekar
The article outlines a vision and recent progress toward a decentralized system that learns from real‑time building data, treating the building itself as a data center to enhance privacy and reduce dependence on external clouds. It introduces a lightweight, Kubernetes‑like orchestration framework for deploying AI services on low‑power microcontrollers, such as those in the Arduino ecosystem, enabling in‑situ learning on sensors. The work also presents experimental results for new online learning algorithms and proposes a cloud‑fog‑edge architecture using KOptim and FIWARE components.
By Christophe C\'erin, Mamadou Sow, Fr\'ed\'eric Andr\`es
The paper presents a training-free diffusion-based motion planner that replaces learned global trajectory scores with analytical local scores derived from obstacle, smoothness, velocity, and inter-agent feasibility terms. By reconstructing trajectory scores through local interactions between neighboring waypoints and nearby constraints, the method decomposes the denoising process while preserving the optimization structure of classical trajectory methods. Experiments demonstrate that this approach generates smooth, feasible trajectories for large multi-agent tasks in complex environments quickly, outperforming learning-based and optimization baselines without requiring training data.
By Michael Y. Fatemi, Jinhao Liang, Ferdinando Fioretto
The paper introduces an agentic forecasting environment built on 2,100+ resolved Polymarket questions, where a language model (Qwen3.5-35B-A3B) learns to gather evidence during rollout via web search, page reading, and financial time series, all filtered to avoid post‑cutoff leaks. Training with single‑epoch GRPO and a Brier‑score reward improves calibration by 30‑40% and reduces search attempts, while the trained policy outperforms four frontier models in evidence‑based forecasting, achieving lower soft‑Brier scores at roughly 5% of the inference cost. The authors release the environment, dataset, and per‑rollout records as a reusable harness for temporal forecasting agents.
By Yusuf Afifi, Artur Kiulian, Anton Polishko, Mykola Khandoga, Hamudi Naanaa, Alina Krasnobrizha
AdaptManip is a fully autonomous framework that enables humanoid robots to navigate, lift, and deliver objects without human demonstrations. It combines a recurrent state estimator for real‑time object tracking, a whole‑body locomotion policy with residual manipulation control, and a LiDAR‑based global position estimator. Trained entirely in simulation with reinforcement learning, the system achieves zero‑shot deployment on real hardware, outperforming imitation‑learning baselines in adaptability and success rate.
By Morgan Byrd, Donghoon Baek, Kartik Garg, Hyunyoung Jung, Daesol Cho, Maks Sorokin, Robert Wright, Sehoon Ha
SkillSpec is a two‑phase framework for evolving natural‑language skills in large language model agents. The first phase, consensus‑gated evolution, generates candidate skills from complementary editing intents and commits updates only when paired evaluations reach consensus on overall improvement and non‑negative aggregate gain. The second phase, representation specialization, uses signals from the optimization trajectory to choose an appropriate flat, graph, or hybrid structure for the skill, improving success rates by an average of 6.89% over SkillOpt across six benchmarks and three target language models.
By Huancheng Chen, Xiaodi Sun, Zhaoqiong Huang, Shenyang Huang Shreya Singhal, Jingwen Lu
The paper investigates online fair allocation of sequential items to agents with heterogeneous preferences, aiming to maximize generalized-mean welfare. In an i.i.d. arrival setting, a pure greedy algorithm achieves near-optimal “~O(1/T)” average regret without needing distributional knowledge. For nonstationary arrivals, the authors show that a single historical sample per distribution suffices to recover the same regret rate, using re-solving algorithms that remain robust to distribution shifts.
By Zongjun Yang, Rachitesh Kumar, Christian Kroer
CF-JEPA is a JEPA-style latent world model that separates the latent space into controllable and uncontrollable subspaces, allowing distractor information to be captured in the uncontrollable region. This factorization prevents latent collapse and maintains performance across 2D and 3D control tasks, even under distracted conditions. The model is validated on a simulated robot task, demonstrating its practical applicability.
By Morgan Byrd, Robert Wright, Sehoon Ha
The article "Where the Agent Development Lifecycle Fits" discusses how to coordinate the development of agent capabilities with the applications they power. It highlights the importance of aligning agent creation processes with the needs of the end‑use cases they support. The piece appears on Towards Data Science and focuses on integrating agent development into broader application workflows.
By Andrew Hinton
The article titled "How to Build a Control Plane for AI Agents" outlines a nine-step process for granting a large language model (LLM) permission to act. It focuses on the practical steps involved in creating a control plane that manages the actions of AI agents. The post was originally published on Towards Data Science.
By Ari Joury, PhD
The paper discusses how large language model (LLM) agents can perform epistemic actions—actions that improve the evidence base before making a decision. It identifies three modes of such actions: acquiring missing evidence, transforming available evidence, and probing the system to generate revealing responses. The authors introduce the concept of epistemic scaffolding, referring to interfaces, tools, and environments that enable and audit these actions, arguing that agent design must explicitly address how decision-ready evidence is produced.
By Yizhi Liu, Balaji Padmanabhan, Siva Viswanathan
arXiv:2610. 01278v1 Announce Type: new Abstract: Alzheimer's disease (AD) diagnosis requires sequential evidence acquisition under heterogeneous test costs and patient burden.
By Ziwen Yu, Ivan Koychev, Elizabeth Coulthard, Ting Zhou, Bolin Chen, Dian Hong, Zinuo You, Yujiao Wang, Anthony Mulholland, Qiang Liu
The paper introduces PACE, a Provenance-Aware Capability Enforcement system designed to secure tool-using large language model agents by mediating every tool call before execution. PACE employs path confinement to limit influence paths and verifies effects against authenticated authority, distinguishing certified execution contracts from evaluated configurations. Experiments on eight agent‑security benchmarks show that the evaluated configuration reduces attack success in most cases while maintaining near‑native utility.
By Fengpeng Li, Qizhou Wang, Yuke Hu, Kemou Li, Jun Liu, Haiwei Wu, Jiantao Zhou, Di Wang
arXiv:2609.34004v2 Announce Type: replace
Abstract: Equity-relevant news evolves through temporally dependent corporate events, making historical information useful only when event continuity, inform...
By Tong Liu, Lanmiao Liu, Xiang Hu
The paper introduces OMAF, an Online MARL framework that uses a one-step Transformer-based flow policy to generate coordinated actions efficiently. By replacing costly iterative sampling with a single-step action generation and a joint optimization scheme that couples softmax Q-value estimation with flow policy objectives, OMAF maintains expressive multimodal behavior while improving training speed. Experiments on 10 tasks from MPE and MAMuJoCo demonstrate up to 3.4× higher returns and 10.5× better sample efficiency compared to baseline methods.
By Zhuoran Li, Yunzhan Li, Xun Wang, Yihan Du, Longbo Huang
Mem++ is a non‑destructive memory framework for large language model agents that records every document in full, along with its date and author, instead of compressing it at write time. At read time, it retrieves only documents up to the requested time and fuses lexical and semantic rankings, leaving the choice of which version to use to the answering model. Evaluations on OrgMemBench show Mem++ outperforms the strongest baseline by 8.0 to 13.1 points and achieves the best overall score with gpt‑4.1‑mini, while also ranking highly on other benchmarks such as LoCoMo and LongMemEval‑S.
By Ahmad Yehia, Aly O. Abdelkareem, Islam Ahmed, Hesham Omran, Khaled Alashmouny, Christian Claudel, Abduallah Mohamed
ESCROW is a post‑deployment maintenance framework for large‑language‑model agents that operate under policy‑bound enterprise workflows. It ensures that any skill revisions proposed by the LLM are only deployed after empirical evaluation, using distributed diagnosis, consensus, per‑category non‑regression guards, cross‑cycle anti‑regression checks, and an accuracy–cost Pareto search to produce a versioned, auditable diff for each change. In a real production financial document‑auditing system, ESCROW achieved the best accuracy–cost trade‑off compared with baseline methods and demonstrated transferability on a public $ au$‑bench.
By Ruoqi Shu, Chen Dan, Xuhui Wang, Tianhua Xu, Mengxi Luo, Yanming Mai, Bo Wan
The paper investigates the gap between intent and behavior in jailbreak attacks on LLM-based robots, showing that many intent jailbreaks fail to produce harmful physical actions due to robot-specific constraints. It introduces POEF, an automated red‑teaming framework that optimizes jailbreak prompts using hidden‑layer gradients and evaluates policy feasibility with a multi‑agent system, achieving an 80% success rate on commercial robots. The authors also propose defense strategies and emphasize the need for stronger countermeasures before widespread deployment of LLM-based robots.
By Xuancun Lu, Zhengxian Huang, Xinfeng Li, Chi Zhang, Xiaoyu Ji, Wenyuan Xu
The paper introduces a reinforcement learning-guided framework for graph transformations aimed at optimizing sparse triangular solve (SpTRSV) kernels. By treating graph transformation as a sequential decision-making task, the RL agent learns matrix-dependent policies that reduce the number of levels by 23% and the coefficient of variation of level costs by 29%, while modifying less than 1% of matrix rows. Experimental results on real-world sparse matrices show significant performance improvements and demonstrate that the learned policies can transfer to unseen matrices via curriculum learning and fine-tuning.
By Buse Y{\i}lmaz