Source code for "Single-Model Uncertainties for Deep Learning", by Natasa Tagasovska and David Lopez-Paz, NeurIPS 2019.
Contents:
├── README.txt (this file)
├── toy (script to generate Figure 1)
├── aleatoric (all experiments related to the aleatoric estimator)
│ └──regression (Section 4.1)
│ ├── regression_experiment.py
│ └── regression_experiment_sweep.json
│ └── joint_estimation.py (Figure 2)
│ └──regression_experiment_parse_table
│ └──causality
│ ├── causal_discovery.py (A.1)
│ └── heterogeneous_qte.py (A.2)
├── epistemic (all experiments related to the epistemic estimator)
│ ├── image_experiment.py (Section 4.2)
│ ├── image_experiment_sweep.py
│ ├── resnet.py
│ └── utils.py
└── data
├── UCI_Datasets
├── kdg.csv
├── pairs_an
├── pairs_ls
├── pairs_ls-s
├── pairs_mn-u
└── pairs_tuebingen
The image datasets will be downloaded once the scripts are run.
This source code is released under a Attribution-NonCommercial 4.0 International license, find out more about it here.