Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Scalable Bayes via Barycenter in Wasserstein Space
Sanvesh Srivastava, Cheng Li, David B. Dunson
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 WASP instead of CMC for variances | consensus Monte Carlo · Linear mixed effects, r=6, k=20 | accuracy: 0.80 -> 0.97 | Improved |
| increase number of subsets k from 10 to 20 (ablation) | CMC · Linear mixed effects variances, r=6 | accuracy: 0.89 -> 0.80 | Hurt |
| increase number of subsets k from 10 to 20 (ablation) | SDP · Linear mixed effects variances, r=6 | accuracy: 0.84 -> 0.77 | Hurt |
| increase number of subsets k from 10 to 20 (ablation) | WASP · Linear mixed effects variances, r=6 | accuracy: 0.97 -> 0.97 | No clear effect |
| increase random effects dimension r from 3 to 6 (ablation) | ADVI · Linear mixed effects variances | accuracy: 0.48 -> 0.09 | Hurt |
| use WASP instead of CMC/SDP with few subsets | CMC and SDP · Probabilistic parafac, k=5 | accuracy: 0.63 -> 0.97 | Improved |
| increase k from 5 to 10 (ablation) | CMC · Probabilistic parafac, pr(x2=1) | accuracy: 0.63 -> 0.96 | Improved |
| use WASP instead of VB | variational Bayes · Gaussian mixture, rho_1 | accuracy: 0.77 -> 0.97 | Improved |
| use WASP on MovieLens variances | CMC · MovieLens, Action variance | accuracy: 0.28 -> 0.92 | Improved |
| use WASP instead of SDP in mixture model combination | SDP · Gaussian mixture, g quantile bands | accuracy | Improved |
| use larger SGLD batch size 2000 -> 4000 (ablation) | SGLD · Linear mixed effects variances | accuracy: 0.68 -> 0.69 (r=3) | No clear effect |
Closest papers by meaning
- Parallel Streaming Wasserstein Barycenters (NeurIPS 2017 · similarity 0.89)
- Fast Bayesian Coresets via Subsampling and Quasi-Newton Refinement (NeurIPS 2022 · similarity 0.87)
- Automated Scalable Bayesian Inference via Hilbert Coresets (JMLR 2019 · similarity 0.87)
- On Markov chain Monte Carlo methods for tall data (JMLR 2017 · similarity 0.86)
- Scalable Bayesian Learning with posteriors (ICLR 2025 · similarity 0.85)
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 2024 · 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.
