NeurIPS 2026 accepted papers on Multi-agent systems and games
415 of the 6,080 papers accepted at NeurIPS 2026 (6.8%) are on Multi-agent systems and games, compared with 3.7% at NeurIPS 2025. They fall into 4 topics.
See NeurIPS 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
LLM agents · 312 papers
- Meta-Cognitive Memory Policy Optimization for Long-Horizon LLM Agents (7 citations)
- Securing AI Agents with Information-Flow Control (3 citations)
- SkillOpt: Executive Strategy for Self-Evolving Agent Skills (2 citations)
- General Agent Evaluation (1 citation)
- Learning Latency-Aware Orchestration for Multi-Agent Systems (1 citation)
- SENTINEL: A Multi-Level Formal Framework for Safety Evaluation of Foundation Model-based Embodied Agents (1 citation)
- SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks (1 citation)
- $\chi$-Bench: Can AI Agents Automate End-to-End, Long-Horizon, Policy-Rich Healthcare Workflows?
- A Matter of TASTE: Improving Coverage and Difficulty of Agent Benchmarks
- Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security
- Agent MechSuits : Mechanistic Subspace Safety Steering for Multi-Turn CLI Agents
- Agent Meltdowns: The Road to Hell Is Paved with Helpful Agents
and 300 more in the interactive map.
Human-AI and goal-conditioned · 50 papers
- AI Agents Push Humans Out of the Loop (1 citation)
- Attention Trajectories as a Diagnostic Axis for Deep Reinforcement Learning
- Can In-Context Learning Support Intrinsic Curiosity?
- Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics
- Emergence of Physical Intelligence via Controllable Information Production
- Grounding Driving VLA via Inverse Kinematics
- Imagine, Don't Narrate: The Generative Bottleneck in World Models of Interaction Dynamics
- JobBench: Aligning Agent Work With Human Will
- Latent Action Reparameterization for Efficient Agent Inference
- Learning to Drive in New Cities Without Human Demonstrations
- MAPLE: Latent Multi-Agent Play for End-to-End Autonomous Driving
- Mecha-nudges for Machines
and 38 more in the interactive map.
Cooperative multi-agent RL · 30 papers
- COSAC: Counterfactual Credit Assignment in Sequential Cooperative Teams
- Decentralized Ranking Aggregation via Gossip: Convergence and Robustness
- Dr. MAS: Stable Reinforcement Learning for Multi-Agent LLM Systems
- Generalized Intention Modeling in Multi-Agent Reinforcement Learning
- Heterogeneous Agent Collaborative Reinforcement Learning
- One Step is Enough: Multi-Agent Reinforcement Learning Based on One-Step Policy Optimization for Order Dispatch on Ride-Sharing Platforms
- PLATO: Pointer Learner for Agent and Task Openness
- Robust Instruction Compliance in Cooperative Multi-Agent Reinforcement Learning
- The Horizon Threshold in Cooperative Multi-Agent Reward-Free Exploration
- World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning
- A Separation Principle for Cooperative Multi-Agent Reinforcement Learning
- Adaptive Communication Range for Scalable Cooperative Multi-Agent Reinforcement Learning
and 18 more in the interactive map.
Games and equilibria · 23 papers
- A Unified Framework for Uniform-Price Resource Allocation Mechanisms
- Bench-MFG: A Benchmark Suite for Learning in Stationary Mean Field Games
- Differentiable Belief-based Opponent Shaping
- Domination-Avoiding Learning Agents Cannot Collude
- GAE Falls Short in Imperfect-Information Self-Play Reinforcement Learning
- Independent Learning of Nash Equilibria in Partially Observable Markov Potential Games with Decoupled Dynamics
- Learning to Recommend in Unknown Games
- Mitigating Retaliatory Algorithmic Collusion in Repeated Games
- NashDreamer: Model-Based Reinforcement Learning for Zero-Sum Imperfect-Information Games
- Optimizing Social Utility in Sequential Experiments
- Outbidding and Outbluffing Elite Humans: Mastering Liar’s Poker via Self-Play and Reinforcement Learning
- Playing Markov Games Without Observing Payoffs
and 11 more in the interactive map.