Code for Backpropagation Through Agents (BPTA/BPPO, AAAI 2024) and AgentMixer: Multi-Agent Correlated Policy Factorization (AAAI 2025). Unified PyTorch implementations for cooperative multi-agent reinforcement learning.
pytorch multi-agent-reinforcement-learning marl aaai-2024 aaai-2025 bpta ctde backpropagation-through-agents bppo agentmixer policy-factorization autoregressive-policies correlated-policies
-
Updated
Oct 29, 2026 - Python