Research map / Learning theory and algorithms / Prediction and decision losses
Harnessing the power of choices in decision tree learning
Guy Blanc, Jane Lange, Chirag Pabbaraju, Colin Sullivan, Li-Yang Tan, Mo Tiwari
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 |
|---|---|---|---|
| Increase the candidate split count from k to k+1 | Top-k · Constructed distributions | accuracy: \frac1{2}+\eps -> 1-\eps | Improved |
| Increase the candidate split count from k to K | Top-k · Monotone constructed distributions | accuracy: \frac{1}{2} + \eps -> 1 - \eps | Improved |
| Increase k across tested settings (ablation) | Top-k decision trees · UCI benchmark datasets | test accuracy: > 5\% (absolute) | Improved |
| Increase k beyond the point where test accuracy plateaus (ablation) | Top-k decision trees · car and hayes-roth | test accuracy | No clear effect |
| Increase k after its test-accuracy peak (ablation) | Top-k decision trees · tic-tac-toe | test accuracy | Hurt |
| Use Top-k instead of Top-1 (ablation) | Decision trees · taiwanese-bankruptcy, credit-card, and electrical-grid-stability | test accuracy | Hurt |
| Use Top-k instead of Top-1 (ablation) | Decision trees · Most other tested datasets | test accuracy | Improved |
| Increase k from 1 to 16 | Decision trees · FICO, depth 5, up to 1407 features | training scalability: 300 features; 10 minutes -> all 1407 features | Improved |
| Increase k from 1 to 16 | Decision trees · FICO, depth 5, increasing training samples | training scalability: 150 samples -> 1000 samples | Improved |
Closest papers by meaning
- Near-Optimal Decision Trees in a SPLIT Second (ICML 2025 · similarity 0.89)
- Optimal Decision Tree Pruning Revisited: Algorithms and Complexity (ICML 2025 · similarity 0.89)
- Branches: Efficiently Seeking Optimal Sparse Decision Trees via AO* (ICML 2025 · similarity 0.87)
- Learning Small Decision Trees with Few Outliers: A Parameterized Perspective (AAAI 2024 · similarity 0.87)
- SORTeD Rashomon Sets of Sparse Decision Trees: Anytime Enumeration (NeurIPS 2025 · similarity 0.86)
- Generative Model for Decision Trees (AAAI 2024 · similarity 0.86)
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.
