Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Pitfalls of Epistemic Uncertainty Quantification through Loss Minimisation
Viktor Bengs, Eyke Hüllermeier, Willem Waegeman
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 |
|---|---|---|---|
| increase regularisation lambda from 1e-5 to 1e-1 | Level-2 empirical loss minimiser · Binary B(0.5), N=100000 | entropy of level-2 distribution: 0.000 -> 1.575 | Mixed |
| increase regularisation lambda from 1 to 10 | Level-2 empirical loss minimiser · Binary B(0.5), N=10 | entropy of level-2 distribution: 9.615 -> 9.958 | Mixed |
| increase sample size N from 10 to 100000 at lambda=10 | Level-2 empirical loss minimiser · Binary B(0.5) | entropy of level-2 distribution: 9.958 -> 4.870 | Improved |
| change aleatoric uncertainty from B(0.5) to B(0.05) | Level-2 empirical loss minimiser · Binary, lambda=10, N=100000 | entropy of level-2 distribution: 4.870 vs 4.851 | No clear effect |
| increase lambda for multiclass uniform case N=100000 | Level-2 empirical loss minimiser · Multiclass | entropy of level-2 distribution: 0.000 -> 1.401 | Mixed |
Closest papers by meaning
- Is Epistemic Uncertainty Faithfully Represented by Evidential Deep Learning Methods? (ICML 2024 · similarity 0.92)
- Rethinking Aleatoric and Epistemic Uncertainty (ICML 2025 · similarity 0.91)
- Second-Order Uncertainty Quantification: A Distance-Based Approach (ICML 2024 · similarity 0.89)
- Epistemic Uncertainty Quantification For Pre-Trained Neural Networks (CVPR 2024 · similarity 0.89)
- Evaluating Epistemic Uncertainty: Beyond OOD Detection and Active Learning (NeurIPS 2026 · similarity 0.86)
- Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage? (NeurIPS 2024 · 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.
