In this paper, we study the problem of reinforcement learning in multi-agent systems where communication among agents is limited. We develop a decentralized actor-critic learning framework in which each agent performs several local updates of its policy and value function, where the latter is approximated by a multi-layer neural network, before exchanging information with its neighbors. This local training strategy substantially reduces the communication burden while maintaining coordination across the network. We establish finite-time convergence analysis for the algorithm under Markov-sampling. Specifically, to attain the $\varepsilon$-accurate stationary point, the sample complexity is of order $\mathcal{O}(\varepsilon^{-3})$ and the communication complexity is of order $\mathcal{O}(\varepsilon^{-1}τ^{-1})$, where tau denotes the number of local training steps. We also show how the final error bound depends on the neural network's approximation quality. Numerical experiments in a cooperative control setting illustrate and validate the theoretical findings.

As machine learning models are increasingly fine-tuned on synthetic data, there is a critical risk of subtle misalignments spreading through interconnected AI systems. This paper investigates subliminal corruption, which we define as undesirable traits are transmitted through semantically neutral data, bypassing standard safety checks. While this phenomenon has been identified, a quantitative understanding of its dynamics is missing. To address this gap, we present a systematic study of the scaling laws, thresholds, and mechanisms of subliminal corruption using a teacher-student setup with GPT-2. Our experiments reveal three key findings: (1) subliminal corruption causes behavioral crossover, degrading the model's overall alignment, not just the targeted trait; (2) alignment fails in a sharp phase transition at a critical threshold of poisoned data, rather than degrading gradually; and (3) interpretability analysis shows the corruption mechanism mimics the model's natural fine-tuning process, making it difficult to detect. These results demonstrate a critical vulnerability in AI systems that rely on synthetic data and highlight the need for new safety protocols that can account for latent threats.

Human team tactics emerge from each player's individual perspective and their ability to anticipate, interpret, and adapt to teammates' intentions. While advances in video understanding have improved the modeling of team interactions in sports, most existing work relies on third-person broadcast views and overlooks the synchronous, egocentric nature of multi-agent learning. We introduce X-Ego-CS, a benchmark dataset consisting of 124 hours of gameplay footage from 45 professional-level matches of the popular e-sports game Counter-Strike 2, designed to facilitate research on multi-agent decision-making in complex 3D environments. X-Ego-CS provides cross-egocentric video streams that synchronously capture all players' first-person perspectives along with state-action trajectories. Building on this resource, we propose Cross-Ego Contrastive Learning (CECL), which aligns teammates' egocentric visual streams to foster team-level tactical situational awareness from an individual's perspective. We evaluate CECL on a teammate-opponent location prediction task, demonstrating its effectiveness in enhancing an agent's ability to infer both teammate and opponent positions from a single first-person view using state-of-the-art video encoders. Together, X-Ego-CS and CECL establish a foundation for cross-egocentric multi-agent benchmarking in esports. More broadly, our work positions gameplay understanding as a testbed for multi-agent modeling and tactical learning, with implications for spatiotemporal reasoning and human-AI teaming in both virtual and real-world domains. Code and dataset are available at https://github.com/HATS-ICT/x-ego.

Domestic AI agents faces ethical, autonomy, and inclusion challenges, particularly for overlooked groups like children, elderly, and Neurodivergent users. We present the Plural Voices Model (PVM), a novel single-agent framework that dynamically negotiates multi-user needs through real-time value alignment, leveraging diverse public datasets on mental health, eldercare, education, and moral reasoning. Using human+synthetic curriculum design with fairness-aware scenarios and ethical enhancements, PVM identifies core values, conflicts, and accessibility requirements to inform inclusive principles. Our privacy-focused prototype features adaptive safety scaffolds, tailored interactions (e.g., step-by-step guidance for Neurodivergent users, simple wording for children), and equitable conflict resolution. In preliminary evaluations, PVM outperforms multi-agent baselines in compliance (76% vs. 70%), fairness (90% vs. 85%), safety-violation rate (0% vs. 7%), and latency. Design innovations, including video guidance, autonomy sliders, family hubs, and adaptive safety dashboards, demonstrate new directions for ethical and inclusive domestic AI, for building user-centered agentic systems in plural domestic contexts. Our Codes and Model are been open sourced, available for reproduction: https://github.com/zade90/Agora

We study the selection of agents based on mutual nominations, a theoretical problem with many applications from committee selection to AI alignment. As agents both select and are selected, they may be incentivized to misrepresent their true opinion about the eligibility of others to influence their own chances of selection. Impartial mechanisms circumvent this issue by guaranteeing that the selection of an agent is independent of the nominations cast by that agent. Previous research has established strong bounds on the performance of impartial mechanisms, measured by their ability to approximate the number of nominations for the most highly nominated agents. We study to what extent the performance of impartial mechanisms can be improved if they are given a prediction of a set of agents receiving a maximum number of nominations. Specifically, we provide bounds on the consistency and robustness of such mechanisms, where consistency measures the performance of the mechanisms when the prediction is accurate and robustness its performance when the prediction is inaccurate. For the general setting where up to $k$ agents are to be selected and agents nominate any number of other agents, we give a mechanism with consistency $1-O\big(\frac{1}{k}\big)$ and robustness $1-\frac{1}{e}-O\big(\frac{1}{k}\big)$. For the special case of selecting a single agent based on a single nomination per agent, we prove that $1$-consistency can be achieved while guaranteeing $\frac{1}{2}$-robustness. A close comparison with previous results shows that (asymptotically) optimal consistency can be achieved with little to no sacrifice in terms of robustness.

Long-horizon agentic search requires iteratively exploring the web over long trajectories and synthesizing information across many sources, and is the foundation for enabling powerful applications like deep research systems. In this work, we show that popular agentic search frameworks struggle to scale to long trajectories primarily due to context limitations-they accumulate long, noisy content, hit context window and tool budgets, or stop early. Then, we introduce SLIM (Simple Lightweight Information Management), a simple framework that separates retrieval into distinct search and browse tools, and periodically summarizes the trajectory, keeping context concise while enabling longer, more focused searches. On long-horizon tasks, SLIM achieves comparable performance at substantially lower cost and with far fewer tool calls than strong open-source baselines across multiple base models. Specifically, with o3 as the base model, SLIM achieves 56% on BrowseComp and 31% on HLE, outperforming all open-source frameworks by 8 and 4 absolute points, respectively, while incurring 4-6x fewer tool calls. Finally, we release an automated fine-grained trajectory analysis pipeline and error taxonomy for characterizing long-horizon agentic search frameworks; SLIM exhibits fewer hallucinations than prior systems. We hope our analysis framework and simple tool design inform future long-horizon agents.

Multi-agent LLM systems fail to realize parallel speedups due to costly coordination. We present CodeCRDT, an observation-driven coordination pattern where agents coordinate by monitoring a shared state with observable updates and deterministic convergence, rather than explicit message passing. Using Conflict-Free Replicated Data Types (CRDTs), CodeCRDT enables lock-free, conflict-free concurrent code generation with strong eventual consistency. Evaluation across 600 trials (6 tasks, 50 runs per mode) shows both benefits and trade-offs: up to 21.1% speedup on some tasks, up to 39.4% slowdown on others, and 100% convergence with zero merge failures. The study formalizes observation-driven coordination for stochastic LLM agents, revealing semantic conflict rates (5-10%) and quality-performance tradeoffs, and provides empirical characterization of when parallel coordination succeeds versus fails based on task structure.

Active particles are entities that sustain persistent out-of-equilibrium motion by consuming energy. Under certain conditions, they exhibit the tendency to self-organize through coordinated movements, such as swarming via aggregation. While performing non-cooperative foraging tasks, the emergence of such swarming behavior in foragers, exemplifying active particles, has been attributed to the partial observability of the environment, in which the presence of another forager can serve as a proxy signal to indicate the potential presence of a food source or a resource patch. In this paper, we validate this phenomenon by simulating multiple self-propelled foragers as they forage from multiple resource patches in a non-cooperative manner. These foragers operate in a continuous two-dimensional space with stochastic position updates and partial observability. We evolve a shared policy in the form of a continuous-time recurrent neural network that serves as a velocity controller for the foragers. To this end, we use an evolutionary strategy algorithm wherein the different samples of the policy-distribution are evaluated in the same rollout. Then we show that agents are able to learn to adaptively forage in the environment. Next, we show the emergence of swarming in the form of aggregation among the foragers when resource patches are absent. We observe that the strength of this swarming behavior appears to be inversely proportional to the amount of resource stored in the foragers, which supports the risk-sensitive foraging claims. Empirical analysis of the learned controller's hidden states in minimal test runs uncovers their sensitivity to the amount of resource stored in a forager. Clamping these hidden states to represent a lesser amount of resource hastens its learned aggregation behavior.

Inference-time computation is a critical yet challenging paradigm for enhancing the reasoning performance of large language models (LLMs). While existing strategies improve reasoning stability and consistency, they suffer from notable limitations: self-correction often reinforces the model's initial biases, and Multi-Agent Collaboration (MAC) often fails due to the lack of efficient coordination mechanisms, leading to collective errors. Although high-performing verifiers can detect reasoning errors, making them reliable requires substantial training. To address these challenges, we introduce a novel inference-time framework, Adaptive Coopetition (AdCo), in which LLM agents utilize an adaptive, UCB-based "coopetition" mechanism. At each round, agents leverage coarse verifier signals to determine whether to collaborate or compete, and iteratively refine their reasoning based on peer feedback. Without relying on high-performance verifiers, our adaptive strategy achieves significant performance gains on mathematical reasoning benchmarks, yielding a 20% relative improvement over baselines on the more challenging dataset. Our approach remains robust and consistent in terms of accuracy under different sample sizes and configurations. This adaptive, signal-guided "coopetition" framework enhances reasoning robustness by leveraging both model knowledge diversity and reasoning trace measures, while also promoting uncertainty-driven exploration, especially when participants have comparable capabilities. From this perspective, our work offers a fresh lens on inference-time computation and paves the way for more resilient multi-agent LLM systems. Our code is available at: https://github.com/AdCo-Research/adaptive-coopetition.

Inertial fusion energy promises nearly unlimited, clean power if it can be achieved. However, the design and engineering of fusion systems requires controlling and manipulating matter at extreme energies and timescales; the shock physics and radiation transport governing the physical behavior under these conditions are complex requiring the development, calibration, and use of predictive multiphysics codes to navigate the highly nonlinear and multi-faceted design landscape. We hypothesize that artificial intelligence reasoning models can be combined with physics codes and emulators to autonomously design fusion fuel capsules. In this article, we construct a multi-agent system where natural language is utilized to explore the complex physics regimes around fusion energy. The agentic system is capable of executing a high-order multiphysics inertial fusion computational code. We demonstrate the capacity of the multi-agent design assistant to both collaboratively and autonomously manipulate, navigate, and optimize capsule geometry while accounting for high fidelity physics that ultimately achieve simulated ignition via inverse design.