Research map / Probabilistic and scientific modeling / Variational and generative sampling
Equilibrium and non-Equilibrium regimes in the learning of Restricted Boltzmann Machines
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 |
|---|---|---|---|
| increase MCMC steps k used for gradient estimate (10 -> 10^4) (ablation) | RBM Rdm-k · MNIST | equilibrium-regime duration (updates): OOE onset ~10^3 updates (k=10) vs >100k updates (k=10^4) | Helped |
| use CD-10 instead of Rdm-10 (ablation) | RBM · MNIST | sample quality from random init | Hurt |
| use PCD-10 instead of CD-10 (ablation) | RBM · MNIST | sample quality | Helped |
| increase k from PCD-10 to PCD-10^3 (ablation) | RBM PCD · MNIST | sample quality vs Rdm-10 OOE: k=10 -> 10^3 | Helped |
| stop sampling at t_G ~ k in OOE Rdm-10 vs long sampling (ablation) | Rdm-10 · MNIST | sample quality | Helped |
| use log-likelihood alone to judge generation quality | RBM CD-10 · MNIST | log-likelihood vs sample quality | Hurt |
Closest papers by meaning
- Fast training and sampling of Restricted Boltzmann Machines (ICLR 2025 · similarity 0.92)
- Learning with Restricted Boltzmann Machines: Asymptotics of AMP and GD in High Dimensions (NeurIPS 2025 · similarity 0.90)
- Cascade of phase transitions in the training of energy-based models (NeurIPS 2024 · similarity 0.89)
- Spherical Boltzmann machines: a solvable theory of learning and generation in energy-based models (NeurIPS 2026 · similarity 0.87)
- BoltzNCE: Learning likelihoods for Boltzmann Generation with Stochastic Interpolants and Noise Contrastive Estimation (NeurIPS 2025 · similarity 0.86)
- Path Gradients after Flow Matching (NeurIPS 2025 · similarity 0.84)
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.