Research map / Learning theory and algorithms / Sample complexity
Marginal-Nonuniform PAC Learnability
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Introduces marginal-nonuniform PAC learning, a variant where sample complexity may depend on the marginal distribution but must hold uniformly over target concepts, and studies its characterization.
Problem
Understanding how PAC learnability changes when guarantees may depend on the marginal distribution over the domain (rather than on the target concept), as a complement to classical nonuniform learning.
Approach
Proposes marginal-nonuniform learning as the mirror image of nonuniform PAC learning (Benedek and Itai 1994): sample complexity may depend on the marginal distribution but must hold uniformly over all target concepts in the class.
Experimental setup
Builds on: nonuniform PAC learning
Closest papers by meaning
- On the Computational Landscape of Replicable Learning (NeurIPS 2024 · similarity 0.88)
- Learning Distributions from Multiple Data Providers (NeurIPS 2026 · similarity 0.87)
- Scale-Sensitive Shattering: Learnability and Evaluability at Optimal Scale (NeurIPS 2026 · similarity 0.86)
- Surprises in Proper Positive-Only Learning (NeurIPS 2026 · similarity 0.86)
- Collaborative Learning with Different Labeling Functions (ICML 2024 · similarity 0.86)
- Fast Rates for Bandit PAC Multiclass Classification (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.