Research map / Optimization / Implicit bias and generalization
When Does Curriculum Learning Help? A Theoretical Perspective
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.
Develops a theoretical framework for curriculum learning based on biased regularized empirical risk minimization, identifying conditions under which curriculum learning provably helps.
Problem
Curriculum learning is empirically effective for training efficiency and generalization, but its theoretical underpinnings are underexplored.
Approach
Models curriculum learning as biased regularized empirical risk minimization and derives conditions under which curriculum ordering provably improves learning.
Closest papers by meaning
- Curriculum Learning by Transfer Learning: Theory and Experiments with Deep Networks (ICML 2018 · similarity 0.92)
- CurBench: Curriculum Learning Benchmark (ICML 2024 · similarity 0.87)
- CurES: From Gradient Analysis to Efficient Curriculum Learning for Reasoning LLMs (ICLR 2026 · similarity 0.86)
- Learning to Solve Complex Problems via Dataset Decomposition (NeurIPS 2025 · similarity 0.86)
- Dual-Level Curriculum Meta-Learning for Noisy Few-Shot Learning Tasks (AAAI 2024 · similarity 0.86)
- Bandit Guided Submodular Curriculum for Adaptive Subset Selection (NeurIPS 2025 · 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.