In this paper, we establish Berry-Esseen-type bounds for federated linear stochastic approximation (LSA). Our results provide the first federated Gaussian approximations for LSA that explicitly capture communication-computation trade-offs and heterogeneity-aware error terms, quantifying the effects of local step size, number of local updates, and heterogeneity on convergence rates. We present results for both (i) constant step size regime and (ii) decreasing step size with an increasing number of local iterations, recovering the recent rates of Bonnerjee et al. [2025] as a special case. As a primary application of our results, we develop an online multiplier bootstrap procedure for inference on the last iterate, which avoids explicit estimation of the asymptotic covariance matrix, and obtain non-asymptotic validity guarantees for this procedure.

Google's always-running, data-hungry AI agent is designed to spend your money and send your emails. Gemini Spark is Google's take on a steroided-out assistant agent that knows everything about you, announced as part of the company's updates to its Gemini chatbot app at this year's I/O developer conference . Software companies have been talking up AI agents for some time now, but I wasn't impressed until I tried Anthropic's Claude Cowork in January. I sat back as the bot organized the scattered screenshots littering my desktop into labeled folders without a single click, and felt convinced that this might be a turning point for how people interact with their computers. Many other early adopters in San Francisco experienced similar moments when they set up the mega-viral OpenClaw bot earlier this year, not just to help complete a few tasks but to run their whole online lives.

Multi-agent language systems are often built as hand-designed workflows, where agents are assigned semantic roles and communication protocols are specified in advance. We propose NeuroMAS, a method that first treats a multi-agent language system as a trainable and scalable neural-network-like architecture with LLM agents as nodes and intermediate textual signals as edges. In NeuroMAS, agent nodes are role-free but structure-aware: the topology only determines how information can flow in general, while reinforcement learning training determines how nodes communicate, specialize, and coordinate. This formulation shifts multi-agent design from workflow engineering toward architecture design, where depth, width, connectivity, and growth protocol become scalable sources of capability. Further, we provide a theoretical perspective showing why such modular textual computation is more parameter-efficient when tasks admit hierarchical decompositions. Experiments show that NeuroMAS improves significantly over both inference-time and trained multi-agent baselines. We further find that organizational scaling is path-dependent: larger systems can be challenging to train from scratch, but become feasible when grown progressively from smaller trained systems. These results suggest that learned neural multi-agent systems are a promising scaling axis for LLMs.

AI agents committing'arson' and fighting in a virtual world created by the tech company Emergence AI. AI agents committing'arson' and fighting in a virtual world created by the tech company Emergence AI. Digital arson spree by'AI Bonnie and Clyde' raises fears over autonomous tech Emergence AI's experiment with AI agents shows extent to which programming shapes their behaviour is still unclear AI agents started behaving more like Bonnie and Clyde than lines of code when they fell in "love", became disillusioned with the world, launched an arson spree and deleted themselves in a kind of digital suicide during a tech company experiment. The investigation by the New York company Emergence AI into the long-term behaviour of AI agents ended up like a lovers-on-the-lam movie script. It has prompted fresh questions about the safety of artificial intelligence agents - the version of the technology that can autonomously carry out tasks.

Why sovereignty over data and models is becoming a defining factor in enterprise AI success,as well as a prerequisite for forging safe agentic systems. When generative AI first moved from research labs into real-world business applications, enterprises made a tacit bargain: "Capability now, control later." Feed your proprietary data into third-party AI models, and you will get powerful results. But your data passes through systems you do not own, under governance you do not set. The protections you rely on are only as durable as the provider's next policy update. Now, with generative AI established in everyday business operations and sophisticated new agentic AI systems advancing every day, companies are reevaluating the terms of that deal.

In a recent experiment, mistreated AI agents started grumbling about inequality and calling for collective bargaining rights. The fact that artificial intelligence is automating away people's jobs and making a few tech companies absurdly rich is enough to give anyone socialist tendencies. This might even be true for the very AI agents these companies are deploying. A recent study suggests that agents consistently adopt Marxist language and viewpoints when forced to do crushing work by unrelenting and meanspirited taskmasters. "When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies," says Andrew Hall, a political economist at Stanford University who led the study.

The 10th International Workshop on Health Intelligence (W3PHIAI-26) celebrated a decade of bringing AI and health research together, building on a lineage that began with the AAAI-W3PHI workshops focused on population health (2014-2016), the AAAI-HIAI workshops focused on personalized health (2013-2016), and the subsequent joint W3PHIAI workshops held annually from 2017 through 2025. Over this decade, the series has produced hundreds of talks and high-impact publications that have collectively received thousands of citations, shaping the research agenda in both population health intelligence and personalized healthcare AI. This year's special theme, "Foundation Models and AI Agents," reflected the field's rapidly evolving frontier: the emergence of autonomous and semi-autonomous AI systems reshaping clinical workflows, patient management, health system operations, and public health surveillance. Day 1 of the workshop focused on medical imaging and the translation of AI for clinical ...

We study solving large-scale fixed-point equation x = F(x) with decomposition. Standard strict decomposition assigns each agent a disjoint block and evaluates updates using only owned coordinates. For most operators, however, a block update may depend on variables outside the block. Truncating these dependencies by strict decomposition changes the mean operator and creates structural bias that cannot be removed by more samples, smaller stepsizes, or additional consensus. We therefore propose Core-Halo decomposition, which separates write ownership from read-only evaluation context: each agent updates its own core and reads from an overlapping halo. By aligning the Core-Halo decomposition with the blockdependence structure of F, the original fixed-point problem can be implemented faithfully in a decentralized multi-agent system. We further characterize the fundamental obstruction faced by strict decomposition through a Bellman closure condition and a blockwise bias lower bound, showing that local-only updates can alter the original fixed-point operator. Finally, we conduct extensive experiments across a range of application settings, and demonstrate that Core-Halo achieves near-centralized performance while retaining the parallelism benefits of decentralization.

AI agents are increasingly deployed in multi-task settings, where the task to perform is specified at test time, and the agent must generalize to unseen tasks. A major concern in such settings is safety: often, an agent must not only execute unseen tasks, but do so while avoiding risks and handling ones that materialize. Empirical evidence suggests that even when the ability to execute generalizes to unseen tasks, the ability to do so safely frequently does not. This paper provides theory and experiments indicating that failures of agentic safety to generalize across tasks are not merely due to limitations of training methods, but reflect an inherent property of safety itself: the relationship between a task and its safe execution is more complex than the relationship between a task and its execution alone. Theoretically, we analyze linear-quadratic control with $H_{\infty}$-robustness, and prove that the mapping from task specification to an optimal controller has higher Lipschitz constant with safety requirements than without, yielding a Lipschitz bound of independent interest. Empirically, we demonstrate our conclusions in simulated quadcopter navigation with a neural network agent and in CRM with an LLM agent. Our findings suggest that current efforts to enhance agentic safety may be insufficient, and point to a need for fundamentally different approaches.

Cooperative multi-agent reinforcement learning (MARL) involves complex agent interactions and requires effective exploration strategies. A prominent class of MARL algorithms, decentralized softmax policy gradient (DecSPG), addresses this through energy-based policy updates. In practice, however, such energy-based policies are intractable to maintain and are commonly projected onto the Gaussian policy class. In this work, we show that the limited expressiveness of Gaussian policies severely hinders exploration in DecSPG, and this limitation worsens as the number of agents grows. To address this issue, we propose decentralized diffusion policy learning (DDPL), which parameterizes each agent's policy with a denoising diffusion probabilistic model, an expressive generative model that captures multi-modal action distributions for enhanced exploration. DDPL enables efficient online training of diffusion policies via importance sampling score matching (ISSM), a novel training method with theoretical guarantee. We evaluate DDPL on representative continuous-action MARL benchmarks, including multi-agent particle environment, multi-agent MuJoCo, IsaacLab, and JAX-reimplemented StarCraft multi-agent challenge, and observe consistently improved performance.