110 papers
This paper identifies a pervasive "agreement bias" in Multimodal LLM verifiers that causes them to over-validate agent behavior, and proposes a lightweight Self-Grounded Verification (SGV) method that significantly improves failure detection and task completion across web navigation, computer use, and robotics by decoupling prior generation from trajectory evaluation.
The paper introduces CRAwDAD, a dual-agent debate framework that enhances causal inference in reasoning language models by facilitating structured dialogue and adversarial critique between agents, significantly improving accuracy on the CLadder benchmark across all levels of Pearl's causal ladder.
This paper introduces IronEngine, a general AI assistant platform featuring a unified orchestration core and a three-phase pipeline that integrates diverse backends, adaptive memory, and extensive tooling to achieve high task completion rates while separating planning quality from execution capability.
This paper addresses the lack of systematic evaluation in Multi-Agent Deep Reinforcement Learning for C-V2X resource allocation by introducing a disentangled benchmark suite of interference games and diverse datasets to isolate specific challenges, ultimately identifying policy robustness and generalization across vehicular topologies as the primary hurdle and demonstrating the superiority of actor-critic methods over value-based approaches.
This paper proposes a novel level-2 inverse game framework that infers agents' estimates of each other's objectives to address the limitations of traditional level-1 methods in decentralized scenarios, demonstrating through theory and experiments that accounting for these mutual misalignments is crucial for accurately predicting strategic interactions.
The paper introduces LLMGreenRec, a novel multi-agent framework leveraging Large Language Models to deduce user green intents and prioritize sustainable product recommendations, thereby bridging the gap between eco-friendly intentions and actions while minimizing the system's own digital carbon footprint.
This paper introduces the Yokai Learning Environment (YLE), a new open-source benchmark for zero-shot coordination that overcomes the saturation of the Hanabi Learning Environment by requiring agents to track moving cards and reason under ambiguous hints, thereby revealing that current state-of-the-art methods fail to maintain consistent internal models when paired with unseen partners.
This paper introduces GRACE, a unified 2D simulator and benchmark that enables transparent, reproducible comparisons of multi-robot path planning algorithms across grid, roadmap, and continuous environments by standardizing task instantiation, execution, and evaluation protocols.
This paper proposes Natural Language-Based Societies of Mind (NLSOMs), a modular framework where large multimodal neural networks communicate via natural language to solve complex AI tasks more effectively than single models, while also exploring the emerging social, economic, and structural challenges of scaling these heterogeneous societies to include billions of agents.
The paper presents COMIC, a fully automated AI system that generates high-quality, diverse comedic sketch videos by employing a multi-agent framework with specialized roles and LLM-based critics trained on YouTube data to iteratively refine content toward professional standards.
This position paper reframes multi-agent memory as a computer architecture challenge by proposing a three-layer hierarchy and identifying critical protocol gaps, with a specific focus on resolving multi-agent memory consistency as the primary obstacle to building reliable and scalable collaborative systems.
This paper extends Causal Normal Form Games to sequential settings by introducing Sequential Causal Multi-Agent Systems, but its comprehensive theoretical and empirical analysis reveals that, under standard rational assumptions, these causal frameworks offer no welfare advantage over classical Stackelberg equilibrium, thereby highlighting a fundamental incompatibility between rational choice and causal reasoning benefits in current game-theoretic models.
KernelSkill is a multi-agent framework that enhances GPU kernel optimization by replacing opaque LLM heuristics with a knowledge-driven, dual-level memory architecture of expert skills, achieving state-of-the-art speedups and a 100% success rate on KernelBench.
This paper introduces WNumMPC, a hierarchical multi-agent navigation framework that combines a reinforcement learning-based planner and a model-based controller to resolve symmetry-induced deadlocks in dense environments by leveraging topological winding numbers for robust, communication-free coordination.
The paper introduces MARLIN, a decentralized reservoir management framework that combines multi-agent reinforcement learning inspired by starling murmurations with LLM-guided reward shaping to effectively handle environmental uncertainties, significantly improving flood response times and computational efficiency compared to traditional methods.
This paper proposes the Olfactory-Auditory augmented Bug algorithm (OA-Bug) for swarm robots to effectively explore denied environments without GNSS or central processing, demonstrating through simulations and real-world experiments that it achieves significantly higher search coverage (96.93%) compared to existing methods like SGBA.
Using artificial market simulations, this study demonstrates that arbitrage trading between leveraged ETFs and futures acts as a critical liquidity bridge during market crashes, supplying depth and tightness from the stable market to the distressed one to mitigate price declines.
This paper introduces temporally sensitive Alternation (ALT) metrics to reveal that conventional outcome-based evaluations can severely mischaracterize multi-agent coordination, as demonstrated by Q-learning agents in a Battle of the Exes variant that achieve high traditional fairness scores but perform significantly worse than random baselines in actual turn-taking dynamics.
This paper addresses the risk of amplified privacy leakage in sequential multi-agent LLM systems by formalizing compositional leakage through mutual information, deriving a theoretical bound on its propagation, and proposing a privacy-regularized training framework that enforces system-level privacy guarantees rather than relying on local agent constraints alone.
This paper introduces XR-DT, an Extended Reality-enhanced Digital Twin framework that integrates a novel Human-Aware Model Predictive Path Integral (HA-MPPI) controller with an attention-based trajectory prediction model to enable safe, efficient, and interpretable motion planning for mobile robots operating alongside humans.