Research map / Optimization / Optimizers
Stochastic Gradient Geodesic MCMC Methods
Chang Liu, Jun Zhu, Yang Song
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 gSGNHT (adaptive thermostat) instead of SGGMC | spherical admixture model posterior sampling · 20News-different | log-perplexity | Improved |
| use subsampled stochastic gradients (batch) instead of full-batch in SGGMC/gSGNHT (ablation) | SGGMC and gSGNHT for SAM · 20News-different | log-perplexity | Improved |
| use SGGMC/gSGNHT sampler instead of GMC-based samplers | spherical admixture model posterior sampling · 20News-different | log-perplexity | Improved |
| use subsampled stochastic gradients instead of full data in SAM inference (ablation) | gSGNHT for SAM · 150K Wikipedia | log-perplexity | Improved |
| use SGGMC/gSGNHT instead of GMC-apprMH and GMC-bGibbs | spherical admixture model posterior sampling · 150K Wikipedia | log-perplexity | Improved |
| set diffusion-noise covariance V(x) to zero (no gradient noise correction) in SGGMC/gSGNHT for SAM | SGGMC and gSGNHT · SAM experiments | log-perplexity | No clear effect |
Closest papers by meaning
- Geodesic Slice Sampling on the Sphere (JMLR 2025 · similarity 0.90)
- Stochastic Gradient Monomial Gamma Sampler (ICML 2017 · similarity 0.87)
- Stochastic Localization via Iterative Posterior Sampling (ICML 2024 · similarity 0.85)
- Sampling from Binary Quadratic Distributions via Stochastic Localization (ICML 2025 · similarity 0.85)
- Generative Modeling on Manifolds Through Mixture of Riemannian Diffusion Processes (ICML 2024 · similarity 0.85)
- Spherical Flows for Sampling Categorical Data (NeurIPS 2026 · 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.
