Research map / Label-efficient and robust learning / Recognition and long-tail
Disambiguation of Weak Supervision leading to Exponential Convergence rates
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 |
|---|---|---|---|
| Use global disambiguation to recover one label per training example before supervised learning, rather than infimum-loss prediction | Partial-label classification on LIBSVM datasets · LIBSVM classification with synthetic partial labels | testing error | Helped |
| Use global disambiguation before learning instead of infimum-loss prediction | Interval regression · Interval regression | prediction regularity | Helped |
| Use the disambiguation framework with input-based weighting | Semi-supervised learning · Concentric-circle, four-class instance | class reconstruction | Helped |
| Use nearest-neighbor weights and disambiguation under the stated learnability assumptions | Proposed disambiguation algorithm · Partial labelling under non-ambiguity, regularity, and minimal-mass assumptions | excess risk: O(n\exp(-n)) | Helped |
| Use the disambiguation framework instead of the infimum-loss baseline | Ranking with partial ordering · Ranking with partial ordering | computation cost | Helped |
Closest papers by meaning
- Learning from weak labelers as constraints (ICLR 2025 · similarity 0.89)
- Label Propagation with Weak Supervision (ICLR 2023 · similarity 0.89)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.88)
- A General Framework for Learning from Weak Supervision (ICML 2024 · similarity 0.86)
- Theoretical Analysis of Weak-to-Strong Generalization (NeurIPS 2024 · similarity 0.86)
- Weak Supervision Performance Evaluation via Partial Identification (NeurIPS 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.