StarCraft II

'StarCraft II' is a complex real-time strategy game used as a benchmark to evaluate cooperative multi-agent reinforcement learning (MARL) algorithms in environments with large joint observation and action spaces.

Papers using StarCraft II (12)

Tleague: A Framework For Competitive Self-play Based Distributed Multi-agent Reinforcement Learning2020 · 12 citesAttention-guided Contrastive Role Representations For Multi-agent Reinforcement Learning2023 · 2 citesBridging MARL to SARL: An Order-Independent Multi-Agent Transformer via Latent Consensus2026Sequence Modeling for N-Agent Ad Hoc Teamwork2025IFS: Information Flow Structure For Multi-agent Ad Hoc System2025Fcmnet: Full Communication Memory Net For Team-level Cooperation In Multi-agent Systems2022 · 6 citesLAGMA: Latent Goal-guided Multi-agent Reinforcement Learning2024 · 1 citesEfficient Communication in Multi-Agent Reinforcement Learning via Variance Based Control2019 · 33 citesLIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning2021 · 3 citesRACA: Relation-Aware Credit Assignment for Ad-Hoc Cooperation in Multi-Agent Deep Reinforcement Learning2022 · 2 citesHeterogeneous Multi-Agent Reinforcement Learning via Mirror Descent Policy Optimization2023FP3O: Enabling Proximal Policy Optimization in Multi-Agent Cooperation with Parameter-Sharing Versatility2023