Agentick is a unified benchmark for sequential decision‑making agents that evaluates RL, LLM, VLM, hybrid, and human agents on 37 procedurally generated tasks across six capability categories, four difficulty levels, and five observation modalities via a single Gymnasium‑compatible interface. It includes a Coding API, oracle reference policies, pre‑built SFT datasets, a composable agent harness, and a live leaderboard. An evaluation of 27 configurations and over 90,000 episodes shows no single approach dominates, with GPT‑5 mini leading overall, PPO excelling in planning and multi‑agent tasks, and the reasoning harness boosting LLM performance by 3‑10×, while ASCII observations outperform natural language.
By Roger Creus Castanyer, Pablo Samuel Castro, Glen Berseth
CODESKILL is an LLM-based framework that learns to extract, evolve, and maintain procedural skills from coding-agent trajectories. It treats skill extraction and skill-bank management as a learnable policy trained with reinforcement learning, using a hybrid reward combining rubric-based skill quality and verifiable execution feedback. Experiments on EnvBench, SWE-Bench Verified, and Terminal-Bench 2 demonstrate that CODESKILL raises average pass rates by 11.03 over a no-skill baseline and by 5.10 over the strongest prompt-based or memory baseline while keeping a compact skill bank.
By Yanzhou Li, Yiran Zhang, Xiaoyu Zhang, Xiaoxia Liu, Yang Liu
The paper introduces an agentic AI-driven immersive simulation for High Dose Rate vaginal cylinder brachytherapy, combining Virtual Reality, mobile computing, and a knowledge-aware assistant powered by Retrieval-Augmented Generation. The system offers a high‑fidelity, risk‑free training environment that provides natural language interfaces and real‑time guidance during complex medical maneuvers. A prototype using Meta Quest 3 and a GPU‑accelerated backend demonstrates acceptable latency and high context precision, answer completeness, and relevance in the pedagogical support.
By Ronghua Xu, Kepha Barasa, Manoj Kumal, Xinyun Liu, Weihua Zhou, Xin Qian
BabelCoder is an agentic framework for automatic code translation that splits the task into specialized agents for translation, testing, and refinement. Each agent focuses on a specific aspect—generating code, validating correctness, or repairing errors—allowing collaborative improvement of translation quality. Evaluated on four benchmark datasets, BabelCoder outperforms four state‑of‑the‑art baselines, achieving an average accuracy of 94.16% and surpassing existing methods in 94% of cases.
By Fazle Rabbi, Soumit Kanti Saha, Tri Minh Triet Pham, Song Wang, Jinqiu Yang
The paper studies a budget‑constrained welfare problem for pooled testing, where agents have heterogeneous utilities and independent probabilities of being healthy. It proves that an optimal dynamic testing policy can achieve at most twice the welfare of the best static overlapping allocation, regardless of population, budget, or pool‑size limit. The authors also identify cases where adaptivity offers no benefit, show that re‑pooling after positive tests is necessary for strict gains, and provide approximation guarantees for greedy algorithms.
By Edwin Lock, Nicholas Lopez, Francisco Marmolejo-Coss\'io, Jose Roberto Tello Ayala, David C. Parkes
VLAA-GUI is a modular framework for autonomous GUI agents that addresses early stopping and repetitive loops by integrating three core components: a Completeness Verifier, a Loop Breaker, and an on-demand Search Agent. The framework also includes optional Coding and Grounding Agents for specialized tasks. Evaluations on five backbones across Linux and Windows benchmarks show strong performance, with some models surpassing human results and the Loop Breaker significantly reducing wasted steps.
By Qijun Han, Haoqin Tu, Zijun Wang, Haoyue Dai, Yiyang Zhou, Nancy Lau, Alvaro A. Cardenas, Yuhui Xu, Ran Xu, Caiming Xiong, Zeyu Zheng, Huaxiu Yao, Yuyin Zhou, Cihang Xie
The paper introduces a topology-driven Multi-Agent Reinforcement Learning (TD-MARL) framework designed to coordinate soft robots in precision manufacturing tasks, specifically to prevent entanglement during unwinding operations in highly constrained environments. By employing centralized learning with a shared topological state, the approach improves observability and training stability, while distributed execution reduces communication demands and enhances system reliability. Simulation results demonstrate that TD-MARL outperforms current deep reinforcement learning methods in convergence speed and anti-winding effectiveness.
By Haoyang Le, Shengxuan Wang, Mohan Chen, Shuo Feng
StraTA introduces Strategic Trajectory Abstraction, a framework that samples a compact strategy from the initial task state and conditions subsequent actions on that strategy, training strategy generation and action execution jointly with a hierarchical GRPO-style rollout design. The method enhances exploration and credit assignment over long horizons by incorporating diverse strategy rollouts and critical self-judgment. Experiments on ALFWorld, WebShop, and SciWorld demonstrate that StraTA consistently improves sample efficiency and final performance, achieving success rates of 93.1% on ALFWorld, 84.2% on WebShop, and a 63.5% overall score on SciWorld, surpassing frontier closed‑source models.
By Xiangyuan Xue, Yifan Zhou, Zidong Wang, Shengji Tang, Philip Torr, Wanli Ouyang, Lei Bai, Zhenfei Yin
Grid‑Orch is a framework that connects Large Language Models (LLMs) to power system simulation via the Model Context Protocol (MCP), allowing engineers to conduct complex distribution grid analyses using natural language. It offers 36 domain‑specific tools across eleven categories—including power flow, voltage analysis, quasi‑static time‑series simulation, and automated optimization—implemented with OpenDSS as the reference engine. The platform supports both cloud‑hosted and locally deployed LLMs, enabling air‑gapped operation, and demonstrates that tasks such as DER interconnection screening can be completed in under two minutes with results identical to traditional scripting.
By Boming Liu, Jin Dong, Jianming Lian
The paper introduces SUN (Semantically UNified) Programs, typed executables that translate grounded relations into optimal control objectives, satisfaction predicates, and learning rewards. Using the Kuafu harness, a foundation model orchestrates scene preparation, verification, residual reinforcement learning, and data generation, repairing candidate programs and calibrating reward weights. Across nine multi‑stage manipulation tasks, Kuafu achieves an 82.03% success rate, outperforms learned baselines, generates demonstrations 10.57× faster than human teleoperation, and transfers zero‑shot to physical Franka and Kinova robots.
By Weiqi Wang, Zhi Li, Yudong Lei, David Martinez, Xiaofeng Gao, Yuxin Jiang, Chenfanfu Jiang, Yingnian Wu, Demetri Terzopoulos, Ran Gong
arXiv:2609.31356v1 Announce Type: new
Abstract: Vision-Language Models (VLMs) exhibit remarkable zero-shot generalization, yet they often encode unwanted or hazardous stylistic domains such as ideali...
By Sumanth Udupa, Mehrtash Harandi, Yadan Luo, Mahsa Baktashmotlagh
arXiv:2609.31234v1 Announce Type: new
Abstract: Problem. Ultra-high-resolution (UHR) remote sensing with vague user intents has two bottlenecks: visual tokens are expensive, and tool calling must be...
By Zhongyu Pang
KuaFu is a unified behavior‑compression layer that reduces each user behavior item to 2–4 tokens, dramatically shrinking per‑item cache size while preserving fidelity through a four‑stage training process. In production across four profiling tasks, it matches or outperforms uncompressed single‑task models, boosts GPU throughput by 37–350%, and saves 190 GPUs. On public benchmarks it consistently beats prior compressors at the same compression ratio, and on RecBench a 4B KuaFu model outperforms its 8B counterpart by 1.90 points, contributing to a 1.37% lift in overall GMV on Tencent’s advertising and recommendation platform.
By Jiahao Hui, Lin Zhu, Yishen Hu, Jingdong Shu, Zetai Jiang, Xining Ran, Ben Tan, Yeshou Cai, Gong Chen, Haijie Gu, Jie Jiang
PIA is a personal intelligence agent that works alongside a consumer health agent to convert health conversations into structured clinical records and to transform those records into a synthesized understanding of the user. It uses a memory system with four controls—extraction, memory, retrieval, and understanding—each supported by a health module that includes a schema, medical alias dictionary, knowledge graph, and temporal rules. The agent demonstrates that deeper memory injection—from simple recall to a health snapshot to a causal trajectory—yields progressively richer answers, while also revealing challenges such as missing self‑reported data, the influence of question phrasing, and the presence of structural noise in causal links.
By Jeonghun Yoon, Dongchan Kim, Hongyeon Yu, Young-Bum Kim, Jaegul Choo
The paper introduces Muslim, an Arabic voice AI platform that delivers grounded Islamic knowledge to users in real time. It combines a NeMo Arabic ASR, an OpenAI-compatible LLM endpoint, and a self-hosted TTS system, supported by a deterministic multi-source retrieval layer across six Model Context Protocol servers. The authors release fine‑tuned Arabic Islamic model artifacts, describe an account‑based metering layer that prevents abuse, and present a three‑layer observability stack that monitors GPU‑bound agent hosts, reporting latency, accuracy, and engineering trade‑offs for production deployment.
By Yahya Mohamed Elnawasany
The Epstein Files Engine is an AI agent developed by the New York Times to help journalists investigate a massive mixed‑media collection released by the U.S. Department of Justice on January 30, 2026, which contains about three million pages of PDFs related to Jeffrey Epstein. The Engine translates reporter questions into Google BigQuery SQL queries across three corpora—Epstein‑related releases, the Times’s archive, and external Epstein‑related news headlines—using an LLM to plan queries and return citation‑rich answers that reporters can verify. Over 100 journalists used the Engine, contributing to at least 20 published stories, and the system includes a Diff method for text‑and‑visual duplicate matching to surface genuinely new information.
By Duy K. Nguyen, Teresa Mondr\'ia Terol, Dylan Freedman, Zach Seward
The paper introduces a prompt‑injection detection framework for email assistants that models attacks as a chain of stages. It combines a text detector, stage‑specific verifiers, rule‑based risk signals, user intent consistency checks, and a logistic decision policy. Experiments on five benchmarks show the framework outperforms pretrained detectors, achieving a mean F1 of 0.406 versus 0.216, and demonstrate that training on benign emails resembling attacks reduces false alarms.
By Ahmad Hashmi, Dhyey Patel, Yunting Yin
MVAgent is a multi‑agent pipeline for multi‑shot video generation that ensures consistent character appearance, stable spatial layout, and continuous character state across shots. The system uses typed conditioning inputs: a Spatial Grounding agent samples camera views, an Observer records shot endings into a continuity memory, a Transition agent builds action and spatial references for subsequent shots, and an Orchestrator composes these inputs into generator requests. Trained with agentic reinforcement learning (Trunk‑GDPO) while keeping the generator and judges frozen, MVAgent achieves the highest cross‑shot consistency and narrative‑planning quality on ViMax‑Bench and is preferred over the strongest agentic baseline in human evaluation.
By Xiangyu Kong, Wenjie Zhou, Fengping Tian, Lihua Fang, Haoqin Sun, Chenyang Lyu, Longyue Wang, Weihua Luo
DAPEVO is a learned visual odometry system that independently estimates image and event correspondences at shared patch locations and fuses their correlation evidence before motion refinement. It maintains image and event descriptors for each tracked patch, using a learned scalar gate to combine modality-specific correlation embeddings for each patch–frame edge, followed by a shared recurrent refinement and bundle‑adjustment update. The method supports event‑only observations and modality‑aware keyframe culling, achieving low trajectory error even when RGB frames are sparse or degraded, outperforming DPVO, RAMP‑VO, and event‑only DEVO on UZH‑FPV and TartanEvent datasets.
By Luca Gandolfi, Simone Nascivera, Roberto Pellerito, Rong Zou, Chiara Plizzari, Davide Scaramuzza
The paper investigates how stable band‑selection methods are and how that stability relates to semantic segmentation performance on the Hyperspectral City V2 dataset. Six band‑selection techniques were tested on ten different class‑balanced ROI sets, producing 60 top‑25 band subsets. Results show that Sim‑LP has the highest intra‑method stability, and together with JMIM+CSNR it also delivers the best segmentation results, achieving up to 2.01 mIoU improvement and 18–22× faster CPU inference for a 9‑band subset, though stability does not consistently predict segmentation quality.
By Jiarong Li, Imad Ali Shah, Enda Ward, Martin Glavin, Edward Jones, Brian Deegan