Research map / Probabilistic and scientific modeling / Bayesian optimization
NOMU: Neural Optimization-based Model Uncertainty
Jakob M Heiss, Jakob Weissteiner, Hanna S Wutte, Sven Seuken, Josef Teichmann
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 NOMU's two-subnetwork architecture and designed uncertainty loss instead of the compared uncertainty estimators | Neural-network model uncertainty estimation · BNN-generated regression test bed | NLPD: BNN1D: -1.65± 0.10 vs -1.08± 0.22; BNN2D: -1.16± 0.05 vs -0.77± 0.07; BNN5D: -0.37± 0.02 vs -0.33± 0.02 | Improved |
| Use NOMU uncertainty bounds as the BO acquisition function | Bayesian optimization · BO average regret 5D | final regret: 2.87e-02 vs 4.14e-02 | Improved |
| Use NOMU uncertainty bounds as the BO acquisition function | Bayesian optimization · BO average regret 10D | final regret: 8.40e-02 vs 6.96e-02 | Mixed |
| Use NOMU uncertainty bounds as the BO acquisition function | Bayesian optimization · BO average regret 20D | final regret: 1.12e-01 vs 1.17e-01 | Improved |
| Use NOMU rather than MCDO for uncertainty estimation | Neural-network regression · UCI Boston | NLPD: 2.68 ± 0.11 vs 2.40 ± 0.07 | Hurt |
| Use NOMU rather than NLM-HPO for uncertainty estimation | Neural-network regression · UCI Energy | NLPD: 0.77 ± 0.06 vs 0.69 ± 0.05 | Hurt |
| Use MC dropout for model uncertainty estimation | Neural-network regression · Noiseless scarce-data toy regression | in-sample uncertainty behavior | Hurt |
| Use HDE for model uncertainty estimation | Neural-network regression · Noiseless scarce-data toy regression | out-of-sample uncertainty behavior | Hurt |
Closest papers by meaning
- Quantifying Point-Prediction Uncertainty in Neural Networks via Residual Estimation with an I/O Kernel (ICLR 2020 · similarity 0.88)
- Distribution-Free Data Uncertainty for Neural Network Regression (ICLR 2025 · similarity 0.87)
- Uncertainty Quantification for Deep Regression using Contextualised Normalizing Flows (NeurIPS 2025 · similarity 0.85)
- A Study of Bayesian Neural Network Surrogates for Bayesian Optimization (ICLR 2024 · similarity 0.85)
- SURE: SUrvey REcipes for building reliable and robust deep networks (CVPR 2024 · similarity 0.84)
- Sketched Lanczos uncertainty score: a low-memory summary of the Fisher information (NeurIPS 2024 · 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.
