Research map / Probabilistic and scientific modeling / Variational and generative sampling
From Monte Carlo to Las Vegas: Improving Restricted Boltzmann Machine Training Through Stopping Sets
Pedro Savarese, Mayank Kakodkar, Bruno Ribeiro
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 |
|---|---|---|---|
| Use LVS-1 instead of CD-10 | RBM with 32 hidden neurons · MNIST testing | average log-likelihood: -153.4 -> -133.3 | Improved |
| Use LVS-1 instead of PCD-10 | RBM with 32 hidden neurons · MNIST testing | average log-likelihood: -138.5 -> -133.3 | Improved |
| Use LVS-1 instead of CD-10 | RBM with 25 hidden neurons · MNIST testing | average log-likelihood: -155.6 -> -137.5 | Improved |
| Use LVS-1 instead of PCD-10 | RBM with 25 hidden neurons · MNIST testing | average log-likelihood: -146.7 -> -137.5 | Improved |
| Vary LVS maximum tour length among K=1, K=3, and K=10 (ablation) | RBM with 32 hidden neurons · MNIST | average log-likelihood: -134.0 (1.0), -133.7 (0.8), -133.3 (1.0) training; -133.3 (1.0), -132.9 (0.7), -132.6 (1.0) testing | No clear effect |
| Vary LVS maximum tour length among K=1, K=3, and K=10 (ablation) | RBM with 25 hidden neurons · MNIST | average log-likelihood: -138.3 (1.3), -138.2 (1.0), -138.1 (1.1) training; -137.5 (1.4), -137.5 (1.1), -137.4 (1.2) testing | No clear effect |
| Use a larger stopping set (ablation) | MCLV/LVS tours on RBM with 32 hidden neurons · MNIST training data, tour-length analysis | probability of finishing in one step | Improved |
| Use learning rates from 10^{-1} to 1 rather than small rates such as 10^{-4} (ablation) | LVS-K on RBM with 25 hidden neurons · MNIST | convergence and performance | Improved |
| Use small learning rates such as 10^{-4} (ablation) | LVS-K on RBM with 25 hidden neurons · MNIST | training speed | Hurt |
Closest papers by meaning
- Fast training and sampling of Restricted Boltzmann Machines (ICLR 2025 · similarity 0.87)
- Learning with Restricted Boltzmann Machines: Asymptotics of AMP and GD in High Dimensions (NeurIPS 2025 · similarity 0.86)
- Ai-sampler: Adversarial Learning of Markov kernels with involutive maps (ICML 2024 · similarity 0.85)
- Microcanonical Langevin Ensembles: Advancing the Sampling of Bayesian Neural Networks (ICLR 2025 · similarity 0.85)
- Markovian Flow Matching: Accelerating MCMC with Continuous Normalizing Flows (NeurIPS 2024 · similarity 0.85)
- AutoStep: Locally adaptive involutive MCMC (ICML 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.
