Research map / Multi-agent systems and games / Games and equilibria
Estimating α-Rank by Maximizing Information Gain
Tabish Rashid, Cheng Zhang, Kamil Ciosek
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 |
|---|---|---|---|
| Select payoff queries by information gain about the α-rank instead of payoff uncertainty | αIG (Bins) · 2 Good, 2 Bad | Bayesian regret | Improved |
| Select payoff queries by information gain about the α-rank instead of payoff uncertainty | αIG (Bins) · 4x4 Gaussian | regret | Improved |
| Use the Wasserstein query objective instead of the entropy-based information-gain objective | αIG · 2 Good, 2 Bad | Bayesian regret | Improved |
| Use the Binning entropy estimator instead of NSB (ablation) | αIG · 3 Good, 5 Bad | Bayesian regret | Improved |
| Use the NSB entropy estimator (ablation) | αIG · 3 Good, 5 Bad | Bayesian regret | Hurt |
| Incorporate payoff antisymmetry and block-wise equality as prior knowledge (ablation) | αIG · 3 Good, 5 Bad | sample efficiency | Improved |
| Incorporate payoff antisymmetry and block-wise equality as prior knowledge (ablation) | αWass · 3 Good, 5 Bad | sample efficiency | Improved |
| Incorporate payoff antisymmetry and block-wise equality as prior knowledge (ablation) | RG-UCB · 3 Good, 5 Bad | sample efficiency | Hurt |
| Increase hallucinated-belief conditioning from once to 100 times (ablation) | αIG · 2 Good, 2 Bad | separation between query objective values | Improved |
| Increase hallucinated-belief conditioning from once to 100 times (ablation) | αWass · 2 Good, 2 Bad | separation between query objective values | Improved |
| Use direct information gain about α-ranks rather than information gain about the payoff distribution | αIG · 2 Good, 2 Bad | relevant-payoff sampling | Improved |
| Use payoff information gain rather than direct α-rank information gain | Payoff · 2 Good, 2 Bad | Bayesian regret | Hurt |
Closest papers by meaning
- Eliciting Kemeny Rankings (AAAI 2024 · similarity 0.86)
- Who’s Gaming the System? A Causally-Motivated Approach for Detecting Strategic Adaptation (NeurIPS 2024 · similarity 0.86)
- Mastering Zero-Shot Interactions in Cooperative and Competitive Simultaneous Games (ICML 2024 · similarity 0.85)
- Is Knowledge Power? On the (Im)possibility of Learning from Strategic Interactions (NeurIPS 2024 · similarity 0.84)
- Decoding Rewards in Competitive Games: Inverse Game Theory with Entropy Regularization (ICML 2025 · similarity 0.84)
- On the Convergence of No-Regret Dynamics in Information Retrieval Games with Proportional Ranking Functions (ICLR 2025 · 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.
