Research map / Label-efficient and robust learning / Recognition and long-tail
Label Efficient Learning by Exploiting Multi-Class Output Codes
Maria Balcan, Travis Dick, Yishay Mansour
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 |
|---|---|---|---|
| Exploit class separation implied by well-separated output-code words by clustering unlabeled data | Linear error correcting output codes · Error correcting output codes | label queries: at most N labels | Improved |
| Use hierarchical single-linkage and labels on a random subset when the class gap is unknown | Error correcting output codes · Error correcting output codes | label queries: t = Õ(αN) | Improved |
| Project one-vs-all data onto the sphere and use robust single-linkage clustering | Linear one-vs-all classifiers on the unit ball · One-vs-all on the unit ball | label queries: at most L labels | Improved |
| Detect separators from locally empty half-balls rather than relying on clustering | Linear output codes under the boundary features condition · Boundary features condition | classification error: error at most ε | Improved |
| Query multiple labels per group and use the majority label | Label-efficient clustering algorithms in the agnostic setting · Agnostic setting | classification error: error at most η + ε | Improved |
| Use a doubling-measure assumption for the error-correcting case | Single-linkage learning · Error correcting output codes | unlabeled sample complexity: n = Õ(d/r_c^{2D} + N/ε^2) | Improved |
Closest papers by meaning
- Labels, Information, and Computation: Efficient Learning Using Sufficient Labels (JMLR 2023 · similarity 0.89)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.87)
- Rebalancing Multi-Label Class-Incremental Learning (AAAI 2025 · similarity 0.86)
- Taming the Sigmoid Bottleneck: Provably Argmaxable Sparse Multi-Label Classification (AAAI 2024 · similarity 0.86)
- Rethinking Self-Distillation: Label Averaging and Enhanced Soft Label Refinement with Partial Labels (ICLR 2025 · similarity 0.86)
- Imprecise Label Learning: A Unified Framework for Learning with Various Imprecise Label Configurations (NeurIPS 2024 · similarity 0.85)
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.
