Research map / Optimization / Nonconvex and smooth optimization
A Stochastic Path Integral Differential EstimatoR Expectation Maximization Algorithm
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 |
|---|---|---|---|
| Use the SPIDER-style control variate based on consecutive iterates instead of the sEM-vr control variate based on the previous outer-loop iterate | Incremental EM · Smooth non-convex finite-sum optimization | optimal K_CE: \epsilon^{-1} n^{2/3} -> \epsilon^{-1} \sqrt{n} | Helped |
| Use SPIDER-EM rather than Online EM | Incremental EM · Smooth non-convex finite-sum optimization | optimal K_CE: \epsilon^{-2} -> \epsilon^{-1} \sqrt{n} | Helped |
| Use SPIDER-EM rather than iEM | Incremental EM · Smooth non-convex finite-sum optimization | optimal K_CE: \epsilon^{-1} n -> \epsilon^{-1} \sqrt{n} | Helped |
| Use SPIDER-EM rather than FIEM with the n^{-2/3} step-size setting | Incremental EM · Smooth non-convex finite-sum optimization | optimal K_CE: \epsilon^{-1} n^{2/3} -> \epsilon^{-1} \sqrt{n} | Helped |
| Use SPIDER-EM rather than FIEM with the n^{-1/3} step-size setting | Incremental EM · Smooth non-convex finite-sum optimization | optimal K_CE: \epsilon^{-3/2} \sqrt{n} vs. \epsilon^{-1} \sqrt{n} | Mixed |
| Use SPIDER-EM rather than Online EM | Incremental EM · MNIST multivariate Gaussian mixture | variability of the estimates | Helped |
Closest papers by meaning
- High-Dimensional Variance-Reduced Stochastic Gradient Expectation-Maximization Algorithm (ICML 2017 · similarity 0.90)
- Fearless Stochasticity in Expectation Propagation (NeurIPS 2024 · similarity 0.86)
- Big Learning Expectation Maximization (AAAI 2024 · similarity 0.85)
- Learning Latent Variable Models via Jarzynski-adjusted Langevin Algorithm (NeurIPS 2025 · similarity 0.85)
- Toward Global Convergence of Gradient EM for Over-Paramterized Gaussian Mixture Models (NeurIPS 2024 · similarity 0.85)
- Learning Mixtures of Experts with EM: A Mirror Descent Perspective (ICML 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.