Research map / Learning theory and algorithms / Prediction and decision losses
All your loss are belong to Bayes
Christian Walder, Richard Nock
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 ISGP (monotone source prior) instead of GP prior for link | Linkgistic · MNIST/Fashion-MNIST generalised linear models | test AUC: 99.1 % -> 99.2 % | Improved |
| learn the link with ISGP-Linkgistic instead of fixed logistic link | logistic regression · Fashion-MNIST | test AUC: 98.5% -> 99.2 % | Improved |
| learn the link with ISGP-Linkgistic instead of fixed logistic link | logistic regression · MNIST | test AUC: 99.9% -> 99.9 % | No clear effect |
| use ISGP-Linkgistic instead of BregmanTron | link-learning GLM · Fashion-MNIST | test AUC: 97.9% -> 99.2 % | Improved |
| use ISGP-Linkgistic instead of SLIsotron | link-learning GLM · MNIST | test AUC: 94.6% -> 99.9 % | Improved |
| use ISGP prior instead of GP prior for isotonic regression | regression model · auto-mpg displacement, Small train | mean test NLL: 871.0 -> 849.4 | Improved |
| use ISGP prior instead of GP prior for isotonic regression | regression model · auto-mpg weight, Small train | mean test NLL: 777.2 -> 770.7 | Improved |
| use ISGP prior instead of GP prior for isotonic regression | regression model · auto-mpg acceleration, Small train | mean test NLL: 2173.3 -> 2216.8 | Hurt |
| use ISGP prior instead of GP prior with Large training set | regression model · auto-mpg weight, Large train | mean test NLL: 735.5 -> 750.7 | Hurt |
| use ISGP prior instead of GP prior with Large training set | regression model · auto-mpg displacement, Large train | mean test NLL: 788.1 -> 780.6 | Improved |
| use PAVA instead of ISGP on small training set | isotonic regression · auto-mpg weight, Small train | mean test NLL: 770.7 -> 835.8 | Hurt |
| use PAVA instead of ISGP on large training set | isotonic regression · auto-mpg power, Large train | mean test NLL: 763.9 -> 755.5 | Improved |
Closest papers by meaning
- Loss Functions and Operators Generated by f-Divergences (ICML 2025 · similarity 0.81)
- Realizable $H$-Consistent and Bayes-Consistent Loss Functions for Learning to Defer (NeurIPS 2024 · similarity 0.81)
- How to Evaluate Behavioral Models (AAAI 2024 · similarity 0.81)
- How to Boost Any Loss Function (NeurIPS 2024 · similarity 0.81)
- Learning with Fitzpatrick Losses (NeurIPS 2024 · similarity 0.80)
- Robust Loss Functions for Training Decision Trees with Noisy Labels (AAAI 2024 · similarity 0.80)
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.
