Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Statistical Model Aggregation via Parameter Matching
Mikhail Yurochkin, Mayank Agarwal, Soumya Ghosh, Kristjan Greenewald, Nghia Hoang
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 |
|---|---|---|---|
| fuse local sparse GPs with SPAHM instead of using local GPs alone | sparse Gaussian process · GSOD temperature, across USA | RMSE: 14.565 -> 8.917 | Improved |
| fuse local sparse GPs with SPAHM instead of using local GPs alone | sparse Gaussian process · GSOD temperature, own state | RMSE: 5.509 -> 5.860 | Hurt |
| use 300 local inducing points instead of 50 (ablation) | local sparse GPs · GSOD temperature, across USA | RMSE: 14.565 -> 15.848 | Hurt |
| fuse local topics with SPAHM instead of Gaussian LDA Gibbs sampler on the full corpus | Gaussian topic model · Gutenberg | UCI coherence: -4.6 -> -2.1 | Improved |
| fuse local topics with SPAHM instead of Gaussian LDA Gibbs sampler | Gaussian topic model · Gutenberg | runtime: 16 hours -> 40 seconds | Improved |
| match with SPAHM instead of k-means matching/pooled k-means | global centroid estimation · simulated Gaussian mixtures | Hausdorff distance | Improved |
| aggregate local HDP-HMMs with SPAHM instead of k-means matching | HDP-HMM on MoCap · CMU MoCap | Hamming distance / ARI | Improved |
| SPAHM matching instead of training a full HDP-HMM jointly | HDP-HMM on MoCap · CMU MoCap | segmentation quality and compute | Mixed |
Closest papers by meaning
- Online Bayesian Moment Matching for Topic Modeling with Unknown Number of Topics (NeurIPS 2016 · similarity 0.86)
- Automated Statistical Model Discovery with Language Models (ICML 2024 · similarity 0.83)
- Bayesian Uncertainty for Gradient Aggregation in Multi-Task Learning (ICML 2024 · similarity 0.83)
- Large Language Bayes (NeurIPS 2025 · similarity 0.82)
- Task Groupings Regularization: Data-Free Meta-Learning with Heterogeneous Pre-trained Models (ICML 2024 · similarity 0.82)
- Learning Mixtures of Gaussian Processes through Random Projection (ICML 2024 · similarity 0.82)
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.
