Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Modeling Uncertainty by Learning a Hierarchy of Deep Neural Connections
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 |
|---|---|---|---|
| Condition discriminative network parameters on the learned generative structure using BRAINet | ResNet-20 last layer · SVHN OOD (CIFAR-10-trained) | AUC-ROC, Ent.: 90.26 -> 92.61 | Helped |
| Replace the last layer with BRAINet and use its OOD-oriented training loss | VGG-13 · TinyImageNet OOD (CIFAR-10-trained) | AUC-ROC: 0.970 -> 0.980 | Helped |
| Replace the last layer with BRAINet and use its OOD-oriented training loss | VGG-13 · TinyImageNet OOD (CIFAR-10-trained) | FPR @TPR=95%: 0.195 -> 0.124 | Helped |
| Use BRAINet instead of standard SGD | VGG-16 · CIFAR-10 | ECE: 0.0483 -> 0.0090 | Helped |
| Weaken the dependence of discriminative parameters on the generative process (ablation) | BRAINet · MNIST | test NLL | Hurt |
| Disconnect the generative-to-discriminative link, yielding independent ensemble networks (ablation) | BRAINet · MNIST | classification accuracy | Hurt |
| Use BRAINet connectivity in place of the standard structure (ablation) | Deep Ensembles · UCI regression benchmarks | RMSE | Mixed |
| Use BRAINet connectivity in place of the standard structure (ablation) | MC-dropout · UCI regression benchmarks | NLL | Mixed |
Closest papers by meaning
- Conservative Uncertainty Estimation By Fitting Prior Networks (ICLR 2020 · similarity 0.91)
- Sparse Uncertainty Representation in Deep Learning with Inducing Weights (NeurIPS 2021 · similarity 0.90)
- Bayesian Uncertainty Estimation for Batch Normalized Deep Networks (ICML 2018 · similarity 0.90)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.88)
- SURE: SUrvey REcipes for building reliable and robust deep networks (CVPR 2024 · similarity 0.88)
- Uncertainty Quantification for Deep Regression using Contextualised Normalizing Flows (NeurIPS 2025 · similarity 0.88)
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.