https://github.com/maximevandegar/neb
Code for reproducing the experiments in the paper Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference.
https://github.com/maximevandegar/neb
empirical-bayes inverse-problems neural-empirical-bayes simulation-based-inference source-distribution-estimation
Last synced: about 1 month ago
JSON representation
Code for reproducing the experiments in the paper Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference.
- Host: GitHub
- URL: https://github.com/maximevandegar/neb
- Owner: MaximeVandegar
- License: mit
- Created: 2020-12-09T10:29:21.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2022-06-21T02:29:10.000Z (almost 3 years ago)
- Last Synced: 2025-03-29T09:41:25.147Z (about 2 months ago)
- Topics: empirical-bayes, inverse-problems, neural-empirical-bayes, simulation-based-inference, source-distribution-estimation
- Language: Python
- Homepage:
- Size: 12 MB
- Stars: 9
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## Getting Started
```commandline
$ pip install -r requirements.txt
```
The implementation provides a simple interface that allows inference in a single line of code.
```python
from inference.neb import McBiasedEstimator, McUnbiasedEstimator, ElboEstimator, IwEstimator
# Define your data, likelihood function, source model, ...
estimator = McBiasedEstimator()
estimator.infer(observations, source_model, optimizer, log_likelihood_fct)
```
The source code for the different estimators was written to be self-contained in a single file for a quick and easy understanding.
[Getting Started](https://github.com/MaximeVandegar/NEB/tree/main/examples)
## Cite
If you make use of this code in your work, please cite our paper:
```
@misc{vandegar2020neural,
title={Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference},
author={Maxime Vandegar and Michael Kagan and Antoine Wehenkel and Gilles Louppe},
year={2020},
}
```