The paper introduces kernel‑managed shared memory, a system‑level abstraction that lets specialized agents write structured, tagged memories while the agent‑system kernel controls retrieval, privacy, and prompt injection. Implemented on AIOS, this design outperforms unmanaged external memory, standard retrieval‑augmented injection, and full context concatenation across GPT‑4o, Llama‑3.1:8B, and Qwen‑2.5:7B, improving personalization scores by 2.4‑4.0 points on a 5‑point scale and reducing latency and token usage by 15‑61%. The results show that centralizing memory management in the kernel delivers most personalization benefits at a fraction of the cost.
By Ryan Lum, Yongfeng Zhang
The paper introduces TRACE, a digital‑advertising diagnostic environment that uses simulated interventions to generate verifiable rewards for training reasoning agents. By injecting controlled interventions into a simulator, the hidden cause of anomalies becomes an oracle label, enabling agents to learn to identify root causes and affected segments through noisy, confounded evidence. Experiments show that reinforcement learning with these synthesized rewards outperforms large prompted baselines, achieving higher accuracy while using fewer tool calls.
By Rui Sun, Zhan Shi, Bing He
JarvisGUI is a new benchmark that tests GUI agents on cross-device workflows involving Android, Windows, and Ubuntu, requiring transfer of intermediate results and coordination across heterogeneous platforms. It formulates tasks as input-output transformations under a lightweight type system, enabling automatic composition of multi-step, cross-device workflows and dynamic evaluation within a unified framework. The benchmark reveals that state-of-the-art open-source GUI agents struggle with state-transfer awareness, cross-platform contextual reasoning, and long-horizon dependency management, exposing a critical capability gap invisible to existing benchmarks.
By Zixiang Chen, Yuheng Lu, Zihao Cheng, Zeming Liu, Jizeng Bai, Ziye Huang, Zhiyin Lin, Zihan Li, Yuhang Guo, Yunhong Wang, Haifeng Wang
The study investigates how speech‑to‑speech (S2S) models handle gender, distinguishing between the acoustic voice and the content’s gender cues. Experiments across five models in English, Spanish, and Mandarin show that while the rendered voice remains unbiased, the models consistently attribute speaker gender based on textual content rather than voice. When content and voice disagree, misgendering rates soar to 90%, whereas agreement yields only 2% misgendering.
By Xiaoqun Liu, Tanu Mitra, Harshit Rajgarhia, Abhishek Mukherji
The paper introduces a taxonomy for resource management systems that span IoT, edge, and cloud layers, extending Wang et al.’s Continuum Orchestration Systems with two new dimensions: the AI Augmentation Paradigm and the Feedback channel. It evaluates six recent architectures and finds that none combine full LLM orchestration with full agent‑layer feedback in a Cloud Continuum setting, highlighting a gap in cross‑tier feedback abstraction. The study emphasizes the need for a unified feedback mechanism to bridge disparate per‑tier signals to the LLM orchestrator.
By Antonino Vaccarella, Lanpei Li, Vincenzo Lomonaco, Massimo Coppola
BRACE introduces an anchored Bellman‑residual correction to address stale critic bias in asynchronous reinforcement learning for language models. By limiting the correction horizon to a prefix of policy tokens and adding a constant‑weight Monte‑Carlo tail, it separates policy correction from reward propagation. The method improves mean@1 on BrowseComp‑Plus by 2.4% and runs 2.46× faster per step than synchronous training while staying stable 50 updates off‑policy.
By Guanqun Zhao, Zijun Xie, Binbin Zheng, Jiafeng Lu, Enlei Gong, Zeyu Chen
Pairit is an online platform that enables researchers to design, test, and deploy live experiments on human-AI collaboration. Using a single YAML configuration file, users can specify an executable experiment graph—including pages, routing, randomization, matchmaking, chat, shared workspaces, server-hosted agents, surveys, timers, and custom HTML components—and combine any number of humans and AI agents in real-time sessions. The platform has been validated through multiple live deployments, including peer-reviewed studies, and captures high-resolution process traces of communication, negotiation, and collaborative work in human-AI dyads.
By Harang Ju, Sinan Aral
The paper titled "With a Thermomix You Lose the Ability to Cook: A Kitchen Machine Analogy for Applications of Generative AI in Education" uses the Thermomix kitchen appliance as an analogy to explore how generative AI tools like ChatGPT are adopted in education. By mapping Thermomix use cases onto learning scenarios and situating them within the ICAP and SAMR frameworks, the authors illustrate how different modes of tool use can either support or undermine meaningful engagement and learning. The study emphasizes that the key issue is not merely whether learners use AI, but how such use shapes their learning processes, offering a conceptual lens for researchers and practitioners to critically examine and guide AI integration in education.
By Nikol Rummel, Valentina Nachtigall, Ernesto Panadero
The paper "Beyond Training: A Feasibility Taxonomy for Inference-Time AI Governance" presents a taxonomy of twenty inference‑time mechanisms for monitoring, verification, and enforcement, each evaluated on a four‑point readiness scale using evidence from four vendors. It applies this taxonomy to a two‑dimensional adversary model and maps the mechanisms to four governance scenarios, finding that most mechanisms are commercially available but only adequate against cooperative or low‑to‑medium‑capability users, not high‑capability state‑level deployers. The study also links inference‑stage controls to hardware‑stage mechanisms through a substitution principle and reports a second‑rater reliability of 0.74.
whyItMatters":"The work identifies the current gaps and readiness of inference‑time governance tools, highlighting that existing mechanisms are insufficient against powerful adversaries and thus informing future regulatory and technical development."
By Samar Ansari
The paper introduces ROTATE, a regret-driven open‑ended training framework that jointly improves an Ad Hoc Teamwork (AHT) agent and an adversarial teammate generator. Unlike traditional two‑stage pipelines, ROTATE alternates between enhancing the agent and generating teammates that specifically probe its collaboration weaknesses. Experiments on Overcooked and Level‑Based Foraging show that ROTATE outperforms existing baselines on unseen teammates, setting a new benchmark for robust, generalizable teamwork.
By Caroline Wang, Arrasy Rahman, Benjamin Nativi, Jiaxun Cui, Yoonchang Sung, Peter Stone
The paper introduces Temporal-Logic-based Causal Diagrams (TL-CDs) for reinforcement learning tasks that involve temporally extended goals. TL-CDs encode causal relationships among environmental properties, complementing deterministic finite automata that model rewards. By leveraging TL-CDs, the authors design an RL algorithm that can predict expected rewards early, leading to significantly reduced exploration and faster convergence to optimal policies.
By Yash Paliwal, Rajarshi Roy, Jean-Rapha\"el Gaglione, Nasim Baharisangari, Daniel Neider, Xiaoming Duan, Ufuk Topcu, Zhe Xu
KairosAgent is an agentic framework that combines a large language model (LLM) reasoner with a time series foundation model (TSFM) forecaster to tackle cross‑domain multimodal time series forecasting. It dynamically invokes analytical tools to improve the LLM’s numerical comprehension and semantic reasoning, then fuses the reasoning outcomes into the TSFM pipeline for more accurate predictions. The approach is further enhanced by a curated large‑scale trajectory corpus and a reinforcement learning paradigm with multi‑turn refinement and turn‑level credit assignment, achieving superior zero‑shot forecasting performance.
By Kun Feng, Ziwei Shan, Yuchen Fang, Yiyang Tan, Sihan Lu, Shuqi Gu, Xingyu Lu, Lintao Ma, Kan Ren
The paper presents a gamified 20Q-style recommender for cybersecurity education that uses reinforcement learning and explainable AI to guide learners through interactive questioning. By acting as a knowledgeable questioner, the system narrows down user-described security scenarios, identifies the underlying threat, and transparently explains its reasoning. The authors detail the system architecture, algorithmic foundations, and provide case studies covering attack vectors such as the Cyber Kill Chain, phishing, ransomware, and web application vulnerabilities.
By Mary Nusrat, Sarfuddin Bhuiyan, Gahangir Hossain
arXiv:2609. 10866v1 Announce Type: new Abstract: Reinforcement learning (RL) agents deployed in real-world environments are often vulnerable to adversarial perturbations in state observations, creating risks in safety-critical applications.
By Tong Li, Saunak Kumar Panda, Yisha Xiang
The paper investigates how synthetic data used for fine‑tuning large language model (LLM) skill routers can lead to catastrophic forgetting of real and out‑of‑distribution (OOD) skills. By evaluating a production router with 34,396 skills, the authors show that synthetic‑data fine‑tuning improves in‑distribution retrieval but degrades performance on real and OOD data. They test several continual‑learning inspired mitigation methods—embedding‑anchor regularization, Learning without Forgetting, Elastic Weight Consolidation, and L2‑initialization—and find that these approaches both preserve OOD retrieval performance and boost synthetic in‑distribution retrieval by up to 13.98% for a 0.6B Qwen retriever and reranker.
By Syed Shariyar Murtaza, Yifan Nie, Utkarsh Soni, Eugene Wen, Arvid Frydenlund
The paper identifies a new failure mode in neurosymbolic systems called Verdict‑Preserving‑Unfaithfulness (VPU), where incorrect formal encodings can still pass solver checks. It introduces Generative Verification (GenV), a method that uses a language model to produce a continuous reference‑equivalence score without relying on explicit localization. Experiments show GenV+HN achieves high AUROC, generalizes to unseen translators, and improves downstream agent performance by 11.3 points.
By Vikash Singh, Debargha Ganguly, Aman Goel, Ali Torkamani, Xiaoxue Han, Joseph Lilien, Ferhat Erata, Vipin Chaudhary
DriftNet is a dual‑head trajectory Transformer designed to detect and localize prompt injection attacks in large language model agents. It processes logged tool‑call trajectories, classifying each as compromised or not while labeling every step as benign, injection point, hijacked, or failed injection. On the AgentDrift benchmark, DriftNet achieves high accuracy, with an F1 score of 0.983, 98.7% exact injection‑point recovery, and low false‑alarm rates.
By Asif Pinjari, Mithun Paul Saint-Germain
The paper introduces EvidenceNet, a runtime assurance layer designed to verify that coordinated AI agent operations achieve an operator’s intended network-wide outcomes across multiple administrative domains. EvidenceNet collects post-change observations from the required authority scopes, checks their freshness and validity, and uses a verifier agent to assess observation content. Experiments on live routing networks demonstrate that this approach can detect successful outcomes that configuration-action logs alone miss, and it rejects completions when observations are sourced incorrectly, substituted, or stale.
By Tianzhu Zhang, Chih-Kai Huang, Meikang Qiu
The paper presents UBCL, a reinforcement learning framework that generates controllable and diverse player behaviors without using human gameplay data. By defining behavior in an N‑dimensional continuous space and training a single PPO‑based multi‑agent policy with target behavior vectors, the method learns how actions affect behavioral statistics such as aggressiveness, mobility, and cooperativeness. Experiments in a custom Unity multiplayer game demonstrate that UBCL achieves greater behavioral diversity than a win‑only baseline and accurately matches specified behavior vectors across a range of targets.
By Atahan Cilan, Atay \"Ozg\"ovde
ProMediConv is a new benchmarking framework for evaluating proactive conversational agents in legal dispute mediation. It models mediation as a multi-stage, party-aware dialogue that incorporates 11 mediation strategies and four party behavior pattern states, and it is built on 972 real-world cases with utterance-level annotations. The framework introduces a fine-grained metric, MAD (Mean Attribute Difference), to capture shifts in party behavior throughout the dialogue, and provides a comprehensive benchmark with diverse models and a tailored baseline, ProMediAgent.
By Zesheng Wei, Mengfan Li, Wenhao Liu, Yixin Zhang, Zilei Wang, Yang Deng