Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Coresets for Scalable Bayesian Logistic Regression
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 sensitivity-weighted logistic regression coresets instead of random subsampling | Bayesian logistic regression · Synthetic datasets | maximum mean discrepancy (MMD): 1-2 orders of magnitude | Helped |
| Use sensitivity-weighted logistic regression coresets instead of random subsampling | Bayesian logistic regression · Synthetic datasets | negative test log-likelihood | Mixed |
| Use logistic regression coresets instead of random subsampling | Bayesian logistic regression · ChemReact | posterior approximation quality | Mixed |
| Use logistic regression coresets instead of random subsampling at larger subset sizes | Bayesian logistic regression · ChemReact | posterior approximation quality | Hurt |
| Use logistic regression coresets instead of random subsampling | Bayesian logistic regression · Webspam | posterior approximation quality | No clear effect |
| Use logistic regression coresets instead of random subsampling | Bayesian logistic regression · Webspam and CovType | MMD | Mixed |
| Use logistic regression coresets instead of random subsampling | Bayesian logistic regression · Webspam, smallest subset size | MMD and test log-likelihood | Hurt |
| Vary the number of clusters or the radius within the tested ranges (ablation) | Logistic regression coreset algorithm · Synthetic and real-world datasets | posterior approximation quality | No clear effect |
| Use original cluster centers instead of adjusted leave-one-out centers (ablation) | Logistic regression coreset construction · Synthetic and real-world datasets | sensitivity: at most a 1% change | No clear effect |
| Use original cluster centers instead of adjusted leave-one-out centers (ablation) | Logistic regression coreset construction · Synthetic and real-world datasets | construction speed: an order of magnitude speed-up | Helped |
Closest papers by meaning
- Fast Bayesian Coresets via Subsampling and Quasi-Newton Refinement (NeurIPS 2022 · similarity 0.92)
- General bounds on the quality of Bayesian coresets (NeurIPS 2024 · similarity 0.91)
- Sparse Variational Inference: Bayesian Coresets from Scratch (NeurIPS 2019 · similarity 0.90)
- Improved Coresets for Vertical Federated Learning: Regularized Linear and Logistic Regressions (ICML 2025 · similarity 0.85)
- The Impact of Coreset Selection on Spurious Correlations and Group Robustness (NeurIPS 2025 · similarity 0.84)
- Bayesian Additive Distribution Regression (NeurIPS 2026 · 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.