Research map / Learning theory and algorithms / Submodular and game algorithms
The Efficiency of the HyperPlay Technique Over Random Sampling
Michael Schofield, Michael Thielscher
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 |
|---|---|---|---|
| apply weight remedy for unbalanced samples (ablation) | HyperPlayer · Battleships In Fog | win rate: 82.0 -> 85.3 | Improved |
| apply weight remedy for unbalanced samples (ablation) | HyperPlayer · Transit | win rate: 72.2 -> 74.6 | Improved |
| apply balance remedy for unbalanced samples (ablation) | HyperPlayer · Mastermind | win rate: 92.8 -> 93.8 | No clear effect |
| apply balance remedy for unbalanced samples (ablation) | HyperPlayer · Hidden Connect | win rate: 37.2 -> 36.0 | No clear effect |
| apply weight remedy for unbalanced samples (ablation) | HyperPlayer · Border Protection | win rate: 51.0 -> 49.2 | Hurt |
| apply either remedy for biased samples (ablation) | HyperPlayer · 8 games | win rate | No clear effect |
| use HyperPlay backtracking sampling instead of random sampling | HyperPlayer · Blind Breakthrough full size, round 32 | states visited per sample | Improved |
| use HyperPlay instead of random sampling over a long game | HyperPlayer · cut-down Blind Breakthrough | cost of sampling relative to random: 2.2% | Improved |
| move from mid-game to end-game sampling | HyperPlay · Transit | median uniformity probability: 0.001 -> 0.999 | Improved |
| move from mid-game to end-game sampling | HyperPlay · Blind Breakthrough | median uniformity probability: 0.156 -> 0.998 | Improved |
| move from mid-game to end-game sampling | HyperPlay · Hidden Connect | median uniformity probability: 0.001 -> 0.001 | No clear effect |
Closest papers by meaning
- The Value of Recall in Extensive-Form Games (AAAI 2025 · similarity 0.86)
- State-Constrained Zero-Sum Differential Games with One-Sided Information (ICML 2024 · similarity 0.85)
- Regret Minimization in Stackelberg Games with Side Information (NeurIPS 2024 · similarity 0.85)
- GUARD: Constructing Realistic Two-Player Matrix and Security Games for Benchmarking Game-Theoretic Algorithms (NeurIPS 2025 · similarity 0.84)
- The impact of uncertainty on regularized learning in games (ICML 2025 · similarity 0.84)
- Local and Adaptive Mirror Descents in Extensive-Form Games (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.
