Research map / Optimization / Optimizers
Efficient preconditioned stochastic gradient descent for estimation in latent variable models
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 |
|---|---|---|---|
| Fisher-preconditioned SGD (Fisher-SGD) instead of MCMC-SAEM | logistic growth NLME model · Logistic growth NLME simulation | RMSE of variance components (Gamma11, Gamma12, Gamma22): Gamma22 RMSE 14.324 vs 16.492 | Helped |
| Fisher-preconditioned SGD instead of MCMC-SAEM | logistic growth NLME model · Logistic growth NLME simulation | empirical 95% coverage of theta: 0.952 vs 0.935 | Helped |
| Fisher-preconditioned SGD instead of MCMC-SAEM | logistic growth NLME model · Logistic growth NLME simulation | RMSE of fixed effects beta1, beta2, alpha: beta1 0.234 vs 0.236; alpha 0.414 vs 0.416 | No clear effect |
| Warm-up procedure (pre-heating, heating, decreasing steps) for learning rate and stabilized preconditioner during pre-heating | Fisher-SGD · SBM simulation | convergence stability | Helped |
| Using a block-diagonal-free full FIM estimate from Fisher-SGD (vs saemix block-diagonal FIM) | confidence intervals in NLME · Logistic growth NLME simulation | coverage of variance parameters: Gamma22 coverage 0.948 vs 0.905 | Helped |
Closest papers by meaning
- High-Dimensional Variance-Reduced Stochastic Gradient Expectation-Maximization Algorithm (ICML 2017 · similarity 0.88)
- On Estimation in Latent Variable Models (ICML 2021 · similarity 0.87)
- Low-Variance Black-Box Gradient Estimates for the Plackett-Luce Distribution (AAAI 2020 · similarity 0.86)
- A Variational Analysis of Stochastic Gradient Algorithms (ICML 2016 · similarity 0.86)
- Learning Latent Variable Models via Jarzynski-adjusted Langevin Algorithm (NeurIPS 2025 · similarity 0.84)
- Identification of Latent Confounders via Investigating the Tensor Ranks of the Nonlinear Observations (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.