Research map / Optimization / Implicit bias and generalization
Feature learning via mean-field Langevin dynamics: classifying sparse parities and beyond
Taiji Suzuki, Denny Wu, Kazusato Oko, Atsushi Nitanda
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 |
|---|---|---|---|
| increase sample size n to Theta(d^2) for 2-sparse parity | Two-layer mean-field NN trained with MFLD · 2-sparse parity, d in 5..150, n in 50..2000 | test accuracy | Improved |
| sample size between linear and quadratic in d for 2-sparse parity | MFLD two-layer network · 2-sparse parity | test accuracy | Mixed |
| raise parity degree to k=3 | MFLD two-layer network · 3-sparse parity | test accuracy | Mixed |
| raise parity degree to k=4 | MFLD two-layer network · 4-sparse parity | test accuracy | Hurt |
Closest papers by meaning
- Learning Multi-Index Models with Neural Networks via Mean-Field Langevin Dynamics (ICLR 2025 · similarity 0.91)
- Improved Particle Approximation Error for Mean Field Neural Networks (NeurIPS 2024 · similarity 0.89)
- Propagation of Chaos for Mean-Field Langevin Dynamics and its Application to Model Ensemble (ICML 2025 · similarity 0.89)
- Mean-Field Langevin Dynamics for Signed Measures via a Bilevel Approach (NeurIPS 2024 · similarity 0.88)
- Mean-field Analysis on Two-layer Neural Networks from a Kernel Perspective (ICML 2024 · similarity 0.88)
- Convergence of Mean-Field Langevin Stochastic Descent-Ascent for Distributional Minimax Optimization (ICML 2025 · similarity 0.87)
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.
