Research map / Learning theory and algorithms / Submodular and game algorithms
Learning from Group Comparisons: Exploiting Higher Order Interactions
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 |
|---|---|---|---|
| model pairwise player interactions with full pairwise matrix (Basic HOI) instead of individual scores (LR) | individual-score logistic regression · HotS Tournament (hero) | accuracy (%): 59.73 -> 80.59 | Helped |
| model pairwise interactions with latent factors (Factorization HOI) instead of individual scores (LR) | HOI model · HotS Tournament (hero) | accuracy (%): 59.73 -> 77.84 | Helped |
| use Factorization HOI instead of LR | individual-score models · Dota 2 (player) | accuracy (%): 65.98 -> 68.25 | Helped |
| use Factorization HOI instead of LR | individual-score models · HotS Public (player) | accuracy (%): 54.01 -> 55.76 | Helped |
| use Factorization HOI instead of LR | individual-score models · HotS Public (hero) | accuracy (%): 54.34 -> 54.75 | Helped |
| use Factorization HOI instead of Basic HOI when many individuals (ablation) | HOI model · HotS Public (player) | accuracy (%): 53.39 -> 55.76 | Helped |
| use Factorization HOI instead of Basic HOI when few individuals (heroes) (ablation) | HOI model · HotS Tournament (hero) | accuracy (%): 80.59 -> 77.84 | Hurt |
| use Basic HOI instead of Factorization HOI on large player set (30,452 players) (ablation) | HOI model · Dota 2 (player) | accuracy (%): oom vs 68.25 | Hurt |
| train Trueskill for 10 epochs instead of 1 | Trueskill · HotS Tournament (hero) | accuracy (%): 62.90 -> 58.48 | No clear effect |
Closest papers by meaning
- Learning the Value of Teamwork to Form Efficient Teams (AAAI 2020 · similarity 0.87)
- EnMatch: Matchmaking for Better Player Engagement via Neural Combinatorial Optimization (AAAI 2024 · similarity 0.86)
- An Analysis of Elo Rating Systems via Markov Chains (NeurIPS 2024 · similarity 0.85)
- Aggregating Quantitative Relative Judgments: From Social Choice to Ranking Prediction (NeurIPS 2024 · similarity 0.84)
- From Average-Iterate to Last-Iterate Convergence in Games: A Reduction and Its Applications (NeurIPS 2025 · similarity 0.83)
- A Unified Perturbation Framework for Analyzing Leaderboard Stability and Manipulation (NeurIPS 2026 · similarity 0.83)
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.