Research map / Reinforcement learning / Visual and representation RL
Efficient Learning for AlphaZero via Path Consistency
Dengwei Zhao, Shikui Tu, Lei Xu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 path consistency loss (lambda) to AlphaZero training | PCZero online learning · 13x13 Hex vs MoHex 2.0 | winning rate: 84.3% to 93.5% (beta=0) | Improved |
| add feature consistency loss (beta) on top of value PC (ablation) | PCZero online learning · 13x13 Hex vs MoHex 2.0 | winning rate: 93.5% to 94.1% | Improved |
| add PC loss to AlphaZero training | online learning on Othello · Othello vs 3-ply Edax | winning rate: 69.5% to 74.8% | Improved |
| add PC loss to AlphaZero training | offline learning on 13x13 Hex · MCTS player vs offline AlphaZero | winning rate: 61.5% (MCTS), 52.1% (greedy) | Improved |
| add PC loss to AlphaZero training | offline learning on Othello and Gomoku · MCTS player vs offline AlphaZero | winning rate: Othello 80.5%, Gomoku 64.0% | Improved |
| add PC loss to offline AlphaZero for greedy players | offline learning on Hex 8x8, 9x9, 13x13, Othello, Gomoku · greedy player vs offline AlphaZero | winning rate: 50.5% to 53.1% | No clear effect |
| add feature consistency loss alone (lambda=0) to offline AlphaZero (ablation) | offline 13x13 Hex · MCTS player vs offline AlphaZero | winning rate: 50.3% | No clear effect |
| add feature consistency (beta=1.0) on top of PC loss (ablation) | offline 13x13 Hex · MCTS player vs offline AlphaZero | winning rate: 61.5% to 70.7% | Improved |
| increase value loss weight gamma to 2.0 or 3.0 instead of PC loss (ablation) | offline AlphaZero on 13x13 Hex · MCTS player vs offline AlphaZero (gamma=1.0) | winning rate: 45.9% (gamma=2.0), 55.0% (gamma=3.0) | No clear effect |
| include MCTS heuristic path (scouted by search) in PC window alongside historical path (ablation) | 8x8 Hex MCTS-PCZero · Elo / tournament vs PCZero | winning rate: 50.0% vs final PCZero; converges faster to higher Elo | Improved |
| use only heuristic path (k=5, l=0) vs only historical path (l=5, k=0) (ablation) | 8x8 Hex MCTS-PCZero window · tournament | winning rate: 57.0% (heuristic-only vs historical-only) | Mixed |
| unbounded heuristic path length (k=inf) vs k=5 in MCTS-PCZero (ablation) | 8x8 and 13x13 Hex MCTS-PCZero · tournament | winning rate: 50.8% vs MCTS-PCZero (l=k=5) on 8x8; 53.0% on 13x13 vs l=k=5 | Improved |
Closest papers by meaning
- Generalized Weighted Path Consistency for Mastering Atari Games (NeurIPS 2023 · similarity 0.90)
- Regret-Guided Search Control for Efficient Learning in AlphaZero (ICLR 2026 · similarity 0.87)
- Scaling Laws for a Multi-Agent Reinforcement Learning Model (ICLR 2023 · similarity 0.86)
- EfficientZero V2: Mastering Discrete and Continuous Control with Limited Data (ICML 2024 · similarity 0.86)
- OptionZero: Planning with Learned Options (ICLR 2025 · similarity 0.84)
- Efficient Reinforcement Learning by Discovering Neural Pathways (NeurIPS 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.
