Research map / Learning theory and algorithms / Prediction and decision losses
More Accurate Learning of k-DNF Reference Classes
Brendan Juba, Hengxuan Li
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 |
|---|---|---|---|
| Use 3-DNF reference class instead of conjunctive anchor | explanation rules · Lending lr | coverage: 10.7 -> 33.6 | Improved |
| Use 3-DNF instead of anchor | explanation rules · Lending gbt | coverage: 9.7 -> 35 | Improved |
| Use 3-DNF instead of anchor | explanation rules · Lending nn | coverage: 7.6 -> 21 | Improved |
| Use 3-DNF instead of anchor | explanation rules · Lending lr | precision: 95.6 -> 82.6 | Hurt |
| Use 3-DNF instead of anchor | explanation rules · Lending gbt | precision: 96.2 -> 82.5 | Hurt |
| Use 3-DNF instead of anchor | explanation rules · Lending nn | precision: 95.6 -> 83.5 | Hurt |
| Use 3-DNF instead of LIME-t | explanation rules · Lending lr | precision: 81 -> 82.6 | Improved |
| Use 3-DNF instead of LIME-t | explanation rules · Lending gbt | precision: 81 -> 82.5 | Improved |
| Use 3-DNF instead of LIME-t | explanation rules · Lending nn | precision: 79.6 -> 83.5 | Improved |
| Use 3-DNF instead of LIME-t | explanation rules · Lending lr | coverage: 21.6 -> 33.6 | Improved |
| Use 3-DNF instead of LIME-t | explanation rules · Lending gbt | coverage: 20.2 -> 35 | Improved |
| Use 3-DNF instead of LIME-t | explanation rules · Lending nn | coverage: 17.3 -> 21 | Improved |
Closest papers by meaning
- Beyond Confidence: Reliable Models Should Also Consider Atypicality (NeurIPS 2023 · similarity 0.85)
- Learning with Average Top-k Loss (NeurIPS 2017 · similarity 0.83)
- Debiased Distillation for Consistency Regularization (AAAI 2025 · similarity 0.83)
- Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View (NeurIPS 2020 · similarity 0.83)
- Estimating Learnability in the Sublinear Data Regime (NeurIPS 2018 · similarity 0.83)
- Knowledge Distillation: Bad Models Can Be Good Role Models (NeurIPS 2022 · similarity 0.82)
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.
