Meltingpot
Emerging4papers using it
2025first seen
The 'Meltingpot' dataset/benchmark is a collection of multi-agent environments designed to evaluate the performance of algorithms in complex, decentralized reinforcement learning settings.
Papers using Meltingpot (4)
- Wonder Wins Ways: Curiosity-Driven Exploration through Multi-Agent Contextual CalibrationGeneralizable Agent Modeling for Agent Collaboration-Competition Adaptation with Multi-Retrieval and Dynamic GenerationTransformer World Model for Sample Efficient Multi-Agent Reinforcement LearningSocialJax: An Evaluation Suite for Multi-agent Reinforcement Learning in Sequential Social Dilemmas