Research map / Probabilistic and scientific modeling / Variational and generative sampling
On Energy-Based Models with Overparametrized Shallow Neural Networks
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 |
|---|---|---|---|
| Train both neuron weights and features in the active regime instead of fixing features in the kernel regime | Shallow-neural-network EBMs · Synthetic planted-neuron distributions (J=1, 2, 4) | test KL divergence | Helped |
| Increase teacher output-weight magnitudes | F1 and F2 EBMs trained with maximum likelihood · Synthetic planted-neuron distributions | test KL divergence | Mixed |
| Train the EBM by minimizing KSD instead of maximum likelihood | Shallow-neural-network EBMs · Synthetic planted-neuron distributions | test KL divergence | Helped |
| Use the KSD metric to assess the difference between active- and kernel-regime models | F1 and F2 EBMs · Synthetic planted-neuron distributions | test KSD | No clear effect |
| Train with F1-Stein discrepancy rather than maximum likelihood | F1 and F2 EBMs · Synthetic planted-neuron distributions | test KL divergence | Helped |
Closest papers by meaning
- Joint Learning of Energy-based Models and their Partition Function (ICML 2025 · similarity 0.89)
- A Theoretical Framework For Overfitting In Energy-based Modeling (ICML 2025 · similarity 0.88)
- Energy-Based Modelling for Discrete and Mixed Data via Heat Equations on Structured Spaces (NeurIPS 2024 · similarity 0.88)
- Improving Adversarial Energy-Based Model via Diffusion Process (ICML 2024 · similarity 0.87)
- Learning Energy-Based Models by Cooperative Diffusion Recovery Likelihood (ICLR 2024 · similarity 0.86)
- Improving Adaptivity via Over-Parameterization in Sequence Models (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.