Research map / Multi-agent systems and games / Games and equilibria
Mastering Complex Control in MOBA Games with Deep Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add action mask (ablation) | Base DiRenJie agent · DiRenJie tournament | time to converge: 80 h -> 65 h | Helped |
| Add action mask (ablation) | Base DiRenJie agent · DiRenjie tournament | win rate vs Base: 50.5% | No clear effect |
| Add target attention (ablation) | Base DiRenjie agent · DiRenjie tournament | win rate vs Base: 75% | Helped |
| Add target attention (ablation) | Base DiRenjie agent · DiRenjie tournament | time to converge: 80 h -> 90 h | Hurt |
| Add LSTM (ablation) | Base DiRenjie agent · DiRenjie tournament | win rate vs Base: 73% | Helped |
| Add LSTM (ablation) | Base DiRenjie agent · DiRenjie tournament | time to converge: 80 h -> 100 h | Hurt |
| Use full rollouts rather than partial rollouts (ablation) | Honor of Kings 1v1 agent · Honor of Kings 1v1 | win rate: 70%~80% | Helped |
| Use random initial frames instead of zero-start (ablation) | Honor of Kings 1v1 agent · Honor of Kings 1v1 | win rate: 40% | Hurt |
| Use random initial frames instead of zero-start (ablation) | Honor of Kings 1v1 agent · Honor of Kings 1v1 | convergence speed: 15% | Helped |
Closest papers by meaning
- Towards Playing Full MOBA Games with Deep Reinforcement Learning (NeurIPS 2020 · similarity 0.93)
- Learning Diverse Policies in MOBA Games via Macro-Goals (NeurIPS 2021 · similarity 0.88)
- Actor-Critic Policy Optimization in a Large-Scale Imperfect-Information Game (ICLR 2022 · similarity 0.87)
- Hierarchical Macro Strategy Model for MOBA Game AI (AAAI 2019 · similarity 0.86)
- FightLadder: A Benchmark for Competitive Multi-Agent Reinforcement Learning (ICML 2024 · similarity 0.85)
- Advancing DRL Agents in Commercial Fighting Games: Training, Integration, and Agent-Human Alignment (ICML 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.