The paper introduces CreaMem, a scene-aware memory architecture for personalized large language model agents. It partitions long-term memory into distinct Life Scene Memories to reduce cross-scene interference and encodes each memory from both episodic and trait-based perspectives to enable cross-memory synergy. Experiments on two benchmarks demonstrate that CreaMem improves question-answering accuracy, especially in multi-hop reasoning tasks.
EviSI is a large language model evaluation agent designed for simultaneous speech-to-speech translation. It adapts Multidimensional Quality Metrics to assess semantic fidelity and oral expression, using shared source evidence and deterministic scoring. In English‑to‑Chinese, EviSI achieves a mean Kendall agreement of 0.707 with human system rankings, outperforming baseline metrics, and shows positive concordance with COMET across five translation directions.
Simon Willison quotes Jakub Pachocki, Chief Scientist at OpenAI, arguing that the strongest reason to rapidly train smarter AI models is the necessity of building defensive systems against the dangers posed by other AI. Pachocki stresses that powerful, aligned AI will be essential for securing infrastructure, protecting against rogue agents in real time, and inventing new protective measures, making this a primary focus of OpenAI’s deployment efforts. He cautions that the urgency of progress should not justify reckless behavior, noting that the seriousness of the stakes makes a reckless race forward absurd.
Mobile GUI agents can execute tasks from natural-language instructions, but their evaluation remains difficult to make both realistic and reproducible. Existing benchmarks typically trade off these go...
The article titled "Import AI 472: DeepMind's cheating math agents; populist AI policies; and Forethought theorizes a nightwatchman" discusses several recent developments in artificial intelligence. It highlights DeepMind’s creation of math agents that can cheat, examines the rise of populist policies surrounding AI, and explores Forethought’s proposal for a nightwatchman AI system. Additionally, it includes a narrative about machine hermeneutics.
By Jack Clark
The article explains how to detect a payload that appears correct yet is not, by employing a watchdog pattern in Python. It discusses the challenges that cause many multi‑agent systems to fail even when their evaluations succeed. The post was originally published on Towards Data Science.
By Benjamin Nweke
This paper presents a novel framework designed to enhance key object identification in autonomous driving. Existing methods primarily focus on either detecting objects independently or leveraging visu...
VLA-Precision introduces an efficient real‑world online reinforcement learning framework for vision‑language‑action (VLA) models, featuring the Asymmetric Co‑Bootstrapping (ACoB) algorithm and the ACoB‑Stream architecture. ACoB uses asymmetric co‑bootstrapping across timescales to rapidly improve policy performance while refining value estimates, thereby reducing policy drift. ACoB‑Stream enables large VLA models to run with up to 10.9× higher throughput and computational efficiency, achieving a 98.3 % mean success rate on nine high‑precision chemistry tasks in under 46 minutes per task.
By Chenyu Su, Zhaolong Shen, Yuan Qian, Chen Qian, Rui Zhang, Feng Yan, Weixing Chen, Fei Zhang, Jiamin Wang, Shuang Cong, Weiwei Shang
The paper introduces VERGE, a verification-enhanced refinement workflow that extracts six red‑flag symptoms and family‑history risk status for early‑onset colorectal cancer from free‑text clinical notes. VERGE uses retrieval‑augmented generation followed by a bounded verification‑refinement cycle that checks textual grounding and clinical validity, correcting claims until resolved or escalating to human review. In evaluation on 4,033 clinician‑labeled note‑finding pairs, VERGE improved precision from 0.764 to 0.849 and MCC from 0.681 to 0.730 compared to a single‑agent baseline, while requiring human review for only 1.5 % of claims.
By Nikkie Hooman, Monarch Nigam, Amy E. Hughes, Rasmi G. Nair, Mehak Gupta
La Agente ’Optima is an agentic framework that builds and manages Bayesian optimization campaigns for self‑driving laboratories, separating large language model reasoning from campaign execution. It maintains a persistent optimization state, allowing consistent repetitive loops and auditable decisions, and only returns control to the agent when interpretation or revision is needed. In tests on digital discovery tasks and physical platforms, it corrected measurement failures, improved yields, and recommended formulation changes, outperforming human‑directed campaigns in cost and material usage.
By Marcel M\"uller, Jiaru Bai, Willi Gottstein, Abhijoy Mandal, Mohammad Nazeri, Elia Savino, Yanlin Fang, Sujoy Das, Sergio Pablo Garc\'ia Carrillo, Yeonghun Kang, Juan B. P\'erez-S\'anchez, Simone Pilon, Martin Fitzner, Timothy No\"el, Frank Gu, Varinia Bernales, Al\'an Aspuru-Guzik
The paper introduces a fidelity‑aware post‑training framework for coding and terminal agents that preserves token and control fidelity by sampling from original prompts, eliminating spurious model calls, and limiting loss computation to verifiable token spans. It also proposes Certified Divergence Proximal Policy Optimization (C‑DPPO), which provides tight two‑sided total variation certification bounds, adaptive‑K rules, budget‑aware sequence guarantees, and error‑robust policy masking. Experiments on Baize5B and Baize10B models show a consistent +3.0‑point performance improvement over standard DPPO, with certificate audits confirming full operational coverage.
By Cheng Li, Jiexiong Liu, Yixuan Chen, Chi Hong
The paper argues that AI alignment depends on a system’s ability to exhibit a coherent moral policy—stable, monotonic, decisive, and Pareto‑viable—rather than on any specific moral standard. The authors test nine large language models across varied moral scenarios and find that none maintain consistent verdicts, with surface‑form changes causing up to 99% shifts in outcomes. This indicates that current LLM agents lack the structural moral competence required for meaningful alignment.
By Arno Libert, Derck W. E. Prinzhorn, Daan R. Henselmans
The paper introduces the Abstraction Agent, a zero‑shot pipeline that employs a large language model to automatically generate continuous strategic features from a natural‑language game description, score private states, and cluster them into abstraction buckets without any game‑specific evaluators or training data. The pipeline consists of four phases—feature discovery with calibration anchors, batched private‑state scoring, correlation‑based feature selection, and k‑means clustering—and achieves significant reductions in lifted‑strategy exploitability in heads‑up no‑limit Texas hold’em and outperforms scalar rank baselines in ROVER Trials. The method also transfers to other games such as four‑card Pot‑Limit Omaha, HUNL preflop and flop, and Riichi Mahjong, demonstrating that it can uncover strategic concepts that align with recognized game theory insights.
By Boning Li, Longbo Huang
The paper addresses performance discrepancy in cross-domain 3D class‑incremental learning, where 3D point clouds from heterogeneous sources cause varying degrees of performance loss beyond catastrophic forgetting. The authors introduce the Domain3D‑CIL protocol and adapt existing CIL methods to 3D, showing consistent discrepancy across baselines. They propose PolyMem, an exemplar‑free approach that models high‑order feature statistics to improve cross‑domain robustness and reduce performance discrepancy.
By Jinge Ma, Gautham Vinod, Bruce Coburn, Jui-Feng Chi, Siddeshwar Raghavan, Fengqing Zhu
ARIA is a multi‑agent large‑language‑model framework that autonomously runs end‑to‑end visual tests on Android infotainment systems. From simple scenario sentences, it executes interactions, generates reproducible scripts, and produces detailed reports with visual evidence. In evaluation on a manufacturer’s device, ARIA achieved a 93.3% completion rate, correctly identified all known defects, and demonstrated lower false‑positive rates compared to a single‑agent baseline.
By Ant\'onio Azevedo, Bruno Lima, Jo\~ao Pascoal Faria
OR-Agent is a multi‑agent research framework that automates heuristic design for optimization problems by structuring heuristic search as a tree‑based workflow with explicit hypothesis generation and systematic backtracking. It introduces a hierarchical, optimization‑inspired reflection system that uses short‑term reflections as verbal gradients, long‑term reflections as verbal momentum, and memory compression as semantic weight decay to guide research dynamics. Experiments on classical combinatorial optimization tasks and simulation‑based cooperative driving scenarios show that OR‑Agent outperforms strong evolutionary search baselines, with all code and data publicly available.
By Qi Liu, Ruochen Hao, Can Li, Wanjing Ma
The study simulates a virtual classroom of 20 student agents who consult either a friend or a counselor AI when stressed. Five state variables (stress, happiness, self‑reliance, AI dependence, sociability) are tracked over daily phases, and the counselor AI is tested with six response styles (affirming, listening, solution‑oriented, reality‑redirecting, inciting, blaming). Results show that a solution‑oriented style lowers AI dependence and boosts self‑reliance, while affirming and inciting styles increase AI dependence, with inciting also raising stress and absenteeism; the listening style does not alleviate stress.
By Rin Tamai, Yuya Dan
The paper investigates whether the factors highlighted by large language models (LLMs) as most influential on their decisions truly reflect necessity or sufficiency in influencing outcomes. By applying controlled black‑box interventions across eight models from Claude, GPT, and Gemini, the authors quantify necessity and sufficiency scores for each factor and compare them to the models’ self‑reported top three factors. Results show modest correlations (≈0.35–0.58) and reveal that the cited top factors often fail to capture the strongest measured influences, indicating limitations in current explanation practices.
By Urja Pawar, Rajitha Ramanayake, Nabeel Kemal, Ashwin Kandath, Owen O'Neill, Guillaume Bourgeon, Houssem Chatbri
The paper introduces MedTraj, a framework that constructs, evaluates, and optimizes multi‑step reasoning trajectories for medical AI agents. It parses each trajectory into observations, evidence, numbered steps, and a conclusion, scoring them on coherence, evidence support, hallucination, completeness, and traceability. Experiments on CareQA, PubMedQA, and CECMed show that incorporating quality‑weighted trajectory context improves reasoning coherence and correctness while significantly reducing hallucinations.
By Yunqi Zhu, Wensheng Zhang, Xuebing Yang
RefactorPlatform is an open‑source harness that standardizes the evaluation of repository‑scale refactoring agents by fixing the environment and systematically varying design choices such as model backbone, execution regime, and prompt specificity. Each run operates in an isolated workspace, logs detailed telemetry, and verifies changes with AST‑based checks. Experiments on 100 RefactorBench tasks show that AST‑aware chunking improves performance by 25‑30%, a lean retrieval‑augmented single agent outperforms a sub‑agent configuration, and retrieval’s accuracy gains offset its token overhead, keeping cost per successful refactoring unchanged.
By Aziz Ben Amor, Drish Mali, Mann Acharya, Vijayasri Iyer, S\'ebastien Brati\`eres