Research map / Probabilistic and scientific modeling / Variational and generative sampling
Flexible mean field variational inference using mixtures of non-overlapping exponential families
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 non-overlapping-mixtures VI scheme instead of naive auxiliary-variable mean field VI | LDpred · LDpred simulations | MSE | Helped |
| Use the non-overlapping-mixtures VI scheme instead of naive auxiliary-variable mean field VI | LDpred · LDpred simulations | correlation | Mixed |
| Use the non-overlapping-mixtures VI scheme instead of MCMC | LDpred · LDpred simulations | correlation | No clear effect |
| Use the non-overlapping-mixtures VI scheme instead of raw observations | LDpred · LDpred simulations | MSE | Helped |
| Set naive-model σ₀² to 1 (ablation) | LDpred naive VI · LDpred simulations | correlation | Hurt |
| Set naive-model σ₀² to 10⁻⁴ rather than 10⁻¹⁰ (ablation) | LDpred naive VI · LDpred simulations | correlation | No clear effect |
| Use the non-overlapping-mixtures scheme instead of naive auxiliary-variable VI | Sparse probabilistic PCA · Sparse PCA simulations | cluster separation | Helped |
| Use the non-overlapping-mixtures scheme instead of naive auxiliary-variable VI | Sparse probabilistic PCA · Sparse PCA simulations | loading sparsity | Helped |
| Reduce naive-model σ₀² toward zero (ablation) | Sparse probabilistic PCA · Sparse PCA simulations | cluster separation | Hurt |
| Use naive sparse probabilistic PCA instead of classical PCA | Sparse probabilistic PCA · Sparse PCA simulations | loading sparsity | Hurt |
| Apply sparsity to LDA topic probabilities using the non-overlapping-mixtures trick | LDA mean-field VI · LDA | sparsity of variational posterior | No clear effect |
Closest papers by meaning
- Fast and Simple Natural-Gradient Variational Inference with Mixture of Exponential-family Approximations (ICML 2019 · similarity 0.89)
- Least squares variational inference (NeurIPS 2025 · similarity 0.89)
- Variational Inference with Mixtures of Isotropic Gaussians (NeurIPS 2025 · similarity 0.88)
- Nearly Dimension-Independent Convergence of Mean-Field Black-Box Variational Inference (NeurIPS 2025 · similarity 0.87)
- Provably Scalable Black-Box Variational Inference with Structured Variational Families (ICML 2024 · similarity 0.87)
- Asymptotics of Alpha-Divergence Variational Inference Algorithms with Exponential Families (NeurIPS 2024 · similarity 0.87)
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.