Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Depth Uncertainty in Neural Networks
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 |
|---|---|---|---|
| optimise ELBO over depth with variational posterior instead of directly maximising marginal likelihood (ablation) | Depth uncertainty network · concrete (5-layer FC) | MLL / posterior collapse | Helped |
| use exact depth posterior instead of variational q(d) (ablation) | DUN · CIFAR10-SVHN OOD rejection | rejection-classification performance | Helped |
| use exact depth posterior instead of variational q(d) (ablation) | DUN · other experiments | performance | No clear effect |
| ensemble networks of different depths (depth-ensembles) instead of standard ensembles | ResNet-50 ensembles · Corrupted CIFAR10 | error and LL | Helped |
| DUN vs SGD point network | ResNet-50 · Flights | RMSE: 34.61 -> 34.69 | No clear effect |
| DUN vs SGD point network | network · Flights | LL: -4.97 -> -4.95 | Helped |
| DUN vs MFVI | network · Flights | TCE: .068 -> .087 | Hurt |
| DUN instead of Dropout | network · UCI gap splits | LL rank | Helped |
| DUN instead of 5-element ensemble | ResNet-50 · Corrupted CIFAR10 | prediction time: 10x faster | Helped |
Closest papers by meaning
- Uncertainty-Aware CNNs for Depth Completion: Uncertainty from Beginning to End (CVPR 2020 · similarity 0.91)
- Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles (NeurIPS 2017 · similarity 0.91)
- Combining Statistical Depth and Fermat Distance for Uncertainty Quantification (NeurIPS 2024 · similarity 0.90)
- SURE: SUrvey REcipes for building reliable and robust deep networks (CVPR 2024 · similarity 0.89)
- Make Me a BNN: A Simple Strategy for Estimating Bayesian Uncertainty from Pre-trained Models (CVPR 2024 · similarity 0.88)
- Uncertainty Quantification for Deep Regression using Contextualised Normalizing Flows (NeurIPS 2025 · similarity 0.87)
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.