Research map / Label-efficient and robust learning / Recognition and long-tail
Querying Partially Labelled Data to Improve a K-nn Classifier
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 |
|---|---|---|---|
| APL querying instead of random querying | 3-nn with partial labels · Random p=0.9 q=0.9, K=3 | average error rate: 61.6(3.73) -> 53.5(1.53) | Helped |
| APL querying instead of random querying | 3-nn with partial labels · Bayes p=0.9 q=0.9, K=3 | average error rate: 62.4(3.33) -> 54.1(1.60) | Helped |
| APL querying instead of random querying | 3-nn with partial labels · Random p=0.7 q=0.5, K=3 | average error rate: 30.8(4.60) -> 27.3(1.67) | Helped |
| APL querying instead of no query | 6-nn with partial labels · Random p=0.9 q=0.9, K=6 | average error rate: 63.3 -> 36.0(1.13) | Helped |
| APL querying instead of max-partial-label (MP) querying | 9-nn with partial labels · Random p=0.9 q=0.9, K=9 | average error rate: 38.3(3.73) -> 31.7(1.13) | Helped |
| ModFF active learning (ACT) instead of random querying | 3-nn with partial labels · Random p=0.7 q=0.5, K=3 | average error rate: 30.8(4.60) -> 32.6(5.53) | Hurt |
| ModFF active learning (ACT) instead of random querying | 9-nn with partial labels · Random p=0.9 q=0.9, K=9 | average error rate: 37.0(2.93) -> 43.4(4.73) | Hurt |
| exact PL scheme instead of approximate APL | 3-nn with partial labels · Random p=0.7 q=0.5, K=3 | average error rate: 27.3(1.67) -> 27.2(1.27) | No clear effect |
| ambiguity-based APL instead of neighbour-based MW | 6-nn with partial labels · Random p=0.9 q=0.9, K=6 | average error rate: 37.8(1.87) -> 36.0(1.13) | Helped |
| ambiguity-based APL instead of neighbour-based MW | 9-nn with partial labels · Random p=0.9 q=0.9, K=9 | average error rate: 33.7(2.33) -> 31.7(1.13) | Helped |
| querying under low partiality with K=9 | all querying schemes · Random p=0.7 q=0.5, K=9 | average error rate: 25.4 -> 24.3(2.40) | No clear effect |
| increase K from 3 to 9 without querying | no query · Random p=0.7 q=0.5 | average error rate: 36.4 -> 25.4 | Helped |
Closest papers by meaning
- Active Classification with Few Queries under Misspecification (NeurIPS 2024 · similarity 0.85)
- Partial-Label Regression (AAAI 2023 · similarity 0.85)
- Reliable Multilabel Classification: Prediction with Partial Abstention (AAAI 2020 · similarity 0.84)
- Extrapolation Towards Imaginary 0-Nearest Neighbour and Its Improved Convergence Rate (NeurIPS 2020 · similarity 0.84)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.84)
- Distilling Reliable Knowledge for Instance-Dependent Partial Label Learning (AAAI 2024 · 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.