Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Correcting Predictions for Approximate Bayesian Inference
Tomasz Kuśmierczyk, Joseph Sakaya, Arto Klami
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 |
|---|---|---|---|
| replace analytic quantile-only decision-maker with neural network decision-maker (ablation) | Proposed decision-maker · last.fm matrix factorization, tilted loss | relative risk reduction: 7-24\% | Improved |
| use neural decision-maker instead of loss-calibrated VI | Standard VI · last.fm, tilted loss | relative risk reduction: 7-24\% | Improved |
| add bootstrap-based regularization prior to neural decision-maker (ablation) | Neural decision-maker · last.fm, squared loss | test risk | Improved |
| add bootstrap-based regularization prior (ablation) | Neural decision-maker · radon multilevel, tilted loss t=0.5 | test risk vs VI iterations | Improved |
| vary regularization parameter lambda (ablation) | Regularized decision-maker · radon, 10^3 iterations | test risk | No clear effect |
| increase sample size S (ablation) | Posterior representation · last.fm | test risk | Improved |
| increase number of quantiles B (ablation) | Posterior representation · last.fm | test risk | Improved |
| apply neural decision-maker with VI failure | q-optimal decision · corn sparse regression | risk: 3 of 10 runs | Improved |
| add regularization in misspecified linear model (ablation) | Neural decision-maker · nonlinear data with linear model | model faithfulness | Improved |
Closest papers by meaning
- Variational Bayesian Decision-making for Continuous Utilities (NeurIPS 2019 · similarity 0.88)
- An Optimization-centric View on Bayes' Rule: Reviewing and Generalizing Variational Inference (JMLR 2022 · similarity 0.88)
- Decision Theoretic Foundations for Conformal Prediction: Optimal Uncertainty Quantification for Risk-Averse Agents (ICML 2025 · similarity 0.86)
- Learning Likelihood-Free Reference Priors (ICML 2025 · similarity 0.86)
- Do Bayesian Neural Networks Actually Behave Like Bayesian Models? (ICML 2025 · similarity 0.86)
- Martingale Posterior Neural Networks for Fast Sequential Decision Making (NeurIPS 2025 · similarity 0.86)
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.
