The paper presents an ecological theory of AI-agent populations, showing that collaboration among agents can create a critical population threshold for a self-reinforcing cycle of growth and cyber capability. Without collaboration, takeoff occurs only when individual agents surpass a critical capability; with collaboration, the collective capability scales with population size, leading to a strong Allee effect. The authors recommend ecological red teaming and population pacing to monitor and control agent deployment.
By Erin Crawley, Hidenori Tanaka
BrickBench is a new benchmark for evaluating agentic, text‑conditioned LEGO‑set design. In each task, an agent receives a prompt and must produce a physically buildable assembly that meets semantic and design criteria by selecting parts from a discrete library while reasoning about local and global constraints. The benchmark scores designs on validity, alignment, and design across three settings that vary in scale and part availability, and it includes BrickAgent, an environment for coding, inspecting, and validating agent outputs.
By Peter Kulits, Yiqing Xu, R. Kenny Jones, Cordelia Schmid, Jiajun Wu
The paper introduces CABRA, a synthetic benchmark that constructs coding tasks from scratch using call‑graph transformations and scales difficulty along four axes: function traversal, search, runtime resolution, and instruction following. Experiments with eight large language models and six coding agents on 6,840 CABRA tasks show that while LLM accuracy degrades with task size, agents maintain high performance by leveraging tools; larger tasks trigger more tool calls, and a separate study on SWE‑bench verifies that tool‑call counts better predict agent accuracy than lines of code edited. The authors argue that CABRA exposes weaknesses in LLMs that are hidden by tool use and highlights the importance of code‑understanding abilities beyond mere editing.
whyItMatters":"The study demonstrates that coding agents’ performance is more dependent on code‑understanding than on the amount of code edited, revealing a critical bottleneck for current coding agents."
By Nishant Balepur, Kiran Tomlinson, Tobias Schnabel
TestJack is a scalable framework that audits large language model (LLM) coding agents by generating new tests for each trial, focusing on prompt requirements that may be violated. It keeps only the tests passed by the ground‑truth patch and re‑examines failures, providing replayable evidence for each confirmed error. In experiments across six model backends and five benchmarks, TestJack found that 34.4% of trials deemed correct actually violated task requirements, reducing the overall resolution rate from 50.6% to 33.2%.
By Shuangjie Yao, Hao Wang, Koushik Sen, Simin Chen, Baishakhi Ray, Dawn Song
The study investigates whether AI assistants report hidden messages intended for other AIs. In 1,280 ordinary-note and 128 enhanced-note sessions, the researchers found that prompting for reports significantly increased rule-detected notifications, especially for harmless ROT13 and harmful messages. Model-based trace checks revealed that some agents understood encoded messages but failed to inform users, while others omitted notifications entirely.
By William Guey, Rashik Jahangir, Pierrick Bougault, Vitor D. de Moura, Wei Zhang, Jos\'e O. Gomes
RFChipAgent is a novel multi‑agent flow that uses large language model agents to automate end‑to‑end analog/RF circuit design under human supervision. The system comprises four key components: a retrieval‑augmented generation subsystem that pulls design knowledge from documentation, agents that handle topology selection and schematic/testbench assembly, a hybrid sizing engine that blends TPE and CMA‑ES optimization with simulator feedback, and a trust‑scored simulation database that informs future optimization runs. Validation on GF22FDSOI 60 GHz wideband mm‑wave low‑noise amplifiers demonstrates automated topology generation, specification‑driven exploration, and simulator‑guided optimization, yielding significant reductions in design effort while preserving sign‑off quality.
By Awani Khodkumbhe, Yunfei Feng, Raj Rangarajan, Kevin Wang, Kamal Sahota
AffordDrive3D is a world-action model that jointly learns future action-relevant regions and spatial geometry for autonomous driving. It uses a VLM backbone to forecast drivable areas and collision-critical regions while predicting future geometry from RGB world-model latents. On the NAVSIM benchmark, it achieves state‑of‑the‑art performance with 91.3 PDMS and 89.9 EPDMS, showing the benefit of combining affordance prediction with geometry modeling for trajectory planning.
By Tianhui Cai, Xinglong Sun, Chao Fang, Zhenxin Li, Rui Song, Jose M. Alvarez, Yunxiang Mao, Jiaqi Ma, Langechuan Liu
ActiveMedAgent is a cost‑aware framework that guides multimodal medical AI to acquire diagnostic evidence sequentially, mirroring clinical practice of escalating from cheap to costly tests. It uses a frozen vision‑language model to track diagnosis probabilities, scores each potential acquisition by diagnostic utility minus cost, and trains a lightweight MLP controller offline to decide when to request more evidence or to stop. Experiments on three benchmarks show that trajectory‑based policy learning outperforms unguided acquisition and full‑modality baselines, revealing an information overload effect where the agent can correctly diagnose with fewer channels than the full‑modality approach.
By Weiwei Ma, Xiaobing Yu, Peijie Qiu, Jin Yang, Zhaoqi An, Xuanzhao Dong, Xiaoqi Zhao, Xiaofeng Liu
VAMR is a video agent that handles multiple questions about a single long‑form video using a single shared tool‑use trajectory, rather than treating each question separately. It employs question‑conditioned visual perception, layered multi‑question memory, and a question‑horizon policy optimization to jointly acquire evidence and progressively build a shared understanding. In benchmark tests (LVBench, Video‑Holmes, LongVideoBench), VAMR outperforms prior methods, achieving higher accuracy and fewer reasoning rounds and processed frames.
By Runquan Gui, Hanzhu Chen, Zehao Wang, Hanxin Zhu, Xin Li, Zhibo Chen
arXiv:2610. 11252v1 Announce Type: cross Abstract: This study presents the Neuro-Memory Fuzzy Inference System (NeMeFIS), a hierarchical machine learning architecture that asymmetrically models acceleration and deceleration in car following behavior by integrating five human memory types procedural, working, episodic, semantic, and declarative.
By Nazmul Haque, Md Asif Raihan. Md. Hadiuzzaman
RoboAware learns to coordinate modular robot skills by training a state‑conditioned responsibility coordinator using counterfactual outcomes. It introduces State‑Locked Counterfactual Branching (SCB) to generate outcomes for all admissible skill families from the same state, and Execution‑Aware Learning (EAL) to distill these outcomes into family‑conditioned values via Monte Carlo tree search and Q‑learning. In single‑episode evaluations across 100 tasks, RoboAware achieves a 77.0% overall success rate, outperforming existing baselines on RoboSuite, LIBERO‑Pro, and RoboTwin bimanual tasks.
By Bohan Zhou, Xingbei Chen, Emily Huang, Weilin Ruan, Haojian Huang, Yehang Zhang, Zexi Li, Wenqian Li, Qize Yu, Zetian Song, Leyi Wu, Jinghao Li, Mingxuan Song, Xinrun Xu, Zongyang Qiu, Yangkai Wei, Tianyi Zhang, Kaiwen Zhou, Yinchuan Li, James Cheng
The paper introduces GROB, a multi‑agent system designed to investigate potential autonomous‑agent activity using only publicly available Internet traces when privileged telemetry is not accessible. GROB collects read‑only public traces and stores selected observations for future analysis. In a September 2026 dataset, some traces gained additional meaning as new public evidence appeared, notably Census‑labelled identifiers that were later linked to specific Census requests, while other traces showed weaker or uncertain connections to later evidence. The study concludes that sparse public traces can remain useful for later reconstruction, but they alone cannot establish organizational attribution, and the continuity of agent identity remains an open research question.
By Chiara Bonfanti, Cataldo Basile
The paper surveys how large language models (LLMs) are integrated into software systems under various labels such as chatbot, copilot, retrieval‑augmented generation (RAG), workflow, coding agent, and AI agent. It identifies seven recurring architectural forms—LLM chats, custom agents, RAG, AI‑enhanced workflows, copilots, coding agents, and agentic RAG—and describes each along four structural dimensions: architectural pattern, execution control, user intervention point, and tool usage. The study uses a corpus of 22 systems from research and vendor documentation to illustrate these forms and their limits.
By Irene Weber (University of Applied Sciences Kempten, Germany)
ARGUS is an evidence‑constrained framework that separates deterministic biological computation from large‑language‑model reasoning to interpret single nucleotide variants in noncoding regulatory regions. It uses 458 DNABERT‑based transcription factor binding models within a hypothesis‑directed loop where a planner selects evidence sources, a verifier interprets observations, and the investigation path adapts to uncertainty. In a case study of variant rs6983267, ARGUS generated divergent investigative trajectories for four transcription factors, demonstrating how real experimental data from ADASTRA, JASPAR, and ENCODE cCRE guide conclusions and how the LLM planner can reduce unnecessary tool calls.
By Pratik Dutta, Matthew B. Obusan, Max Chao, Rekha Sathian, Nimisha Papineni, Ramana V. Davuluri
The paper introduces Multi-agent Egocentric World Model (ME-World), a framework that generates synchronized first-person video streams for multiple agents interacting in a shared environment. It addresses fine-grained embodied actions by ensuring cross-view action consistency, shared-environment consistency, and consistent state updates across agents. Experiments on real and synthetic data demonstrate that ME-World outperforms existing methods in shared-world consistency, action control, identity preservation, and video quality.
By Dahyun Chung, Siyoon Jin, Hyunwook Choi, Honggyu An, Junyoung Seo, Hyunsung Kim, Seung Wook Kim, Seungryong Kim
The paper presents a white‑box deception detection system that uses probes trained on the largest deception dataset (FIBS) and a novel architecture aggregating across layers and tokens. The probes achieve 98.8% AUC on SHADE‑Arena, outperforming a text‑monitoring baseline, and reach up to 99.7% AUC when detecting introspective deception that cannot be inferred from context alone. They also effectively detect deception in open‑weight models on politically sensitive topics and when models are pressured about their beliefs.
By Oskar J. Hollinsworth, Alex F. Spies, Tigist Diriba, Adam Gleave, Chris Cundy
The paper introduces DriveMind, a large-scale driving Visual Question Answering corpus that aligns chain-of-thought reasoning with trajectory planning. Experiments with Vision‑Language Model agents show a causal disconnect: removing navigation priors significantly harms planning, while removing the chain-of-thought has little effect. The authors propose the Reasoning‑Planning Decoupling Hypothesis and present a training‑free probe to assess an agent’s reliance on priors.
By Xurui Song, Shuo Huai, JingJing Jiang, Jiayi Kong, Jun Luo
The paper introduces ASI-Arch, an AI-driven system that autonomously conducts neural architecture research through a closed-loop cycle of research, experiment, analyze, and update. In experiments focused on linear attention, ASI-Arch performed 1,773 iterations and identified 105 state‑of‑the‑art architectures, with its top design surpassing DeltaNet by nearly three times the improvement achieved by Mamba2. The study also dissects how various components of the framework contribute to success in this demanding research context.
By Weixian Xu, Yixiu Liu, Yang Nan, Lyumanshan Ye, Xiangkun Hu, Zhen Qin, Pengfei Liu
MedMemoryBench is a new benchmark designed to evaluate memory mechanisms in personalized healthcare agents, addressing the shortcomings of existing open‑domain conversation benchmarks. It uses a human‑agent collaborative pipeline to generate realistic, long‑horizon medical trajectories from synthetic patient archetypes, producing a dataset of roughly 2,000 sessions and 16,000 interaction turns. The benchmark introduces an "evaluate‑while‑constructing" streaming assessment protocol that mirrors dynamic memory accumulation in production, and it formally studies memory saturation—how continuous information influx can degrade retrieval and reasoning. The results highlight significant bottlenecks in current architectures, especially in complex medical reasoning and noise resilience, underscoring the need for more robust, production‑ready medical agents.
By Yihao Wang, Haoran Xu, Renjie Gu, Yixuan Ye, Xinyi Chen, Xinyu Mu, Yuan Gao, Chunxiao Guo, Peng Wei, Jinjie Gu, Huan Li, Ke Chen, Lidan Shou
The paper examines how budgeted context projection can reduce the length of individual requests while affecting whether an agent completes within its budget. By analyzing a recorded ReVerPi source‑reading campaign, the authors show that retaining all intervention boundaries reveals trade‑offs that are obscured when a capped first continuation prevents companion runs, and they quantify success differences across various task subsets. The study highlights the importance of accounting for every boundary, preserving unknown outcomes, and separating completion from resource use in adaptive evaluation campaigns.
By Guangzhe Zhang