https://github.com/ermongroup/spn_variational_inference
PyTorch implementation for "Probabilistic Circuits for Variational Inference in Discrete Graphical Models", NeurIPS 2020
https://github.com/ermongroup/spn_variational_inference
approximate-inference graphical-models probabilistic-circuits pytorch sum-product-networks variational-inference
Last synced: about 1 month ago
JSON representation
PyTorch implementation for "Probabilistic Circuits for Variational Inference in Discrete Graphical Models", NeurIPS 2020
- Host: GitHub
- URL: https://github.com/ermongroup/spn_variational_inference
- Owner: ermongroup
- Created: 2020-10-22T02:02:15.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2021-10-11T05:19:07.000Z (almost 4 years ago)
- Last Synced: 2025-05-08T01:37:29.402Z (5 months ago)
- Topics: approximate-inference, graphical-models, probabilistic-circuits, pytorch, sum-product-networks, variational-inference
- Language: Python
- Homepage: https://ermongroup.github.io/blog/spnvi/
- Size: 1.45 MB
- Stars: 17
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
This repository contains code for the paper:
[Probabilistic Circuits for Variational Inference in Discrete Graphical Models](https://arxiv.org/abs/2010.11446)
```
"Probabilistic Circuits for Variational Inference in Discrete Graphical Models"
Andy Shih, Stefano Ermon
In Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS), 2020
@inproceedings{ShihEneurips20,
author = {Andy Shih and Stefano Ermon},
title = {Probabilistic Circuits for Variational Inference in Discrete Graphical Models},
booktitle = {Advances in Neural Information Processing Systems 33 (NeurIPS)},
month = {december},
year = {2020},
}
```
Here are commands for running the experiments:
## Ising Models
```
python runising.py --loadgm=1000 --run=123 --mode=2 --n=4
python runising.py --loadgm=1000 --run=123 --mode=1 --n=8
python runising.py --loadgm=1000 --run=123 --mode=1 --n=16
python runising.py --loadgm=1000 --run=123 --mode=1 --n=32
```
, and repeat with loadgm=[1001,1002,1003].
## UAI Inference Competition
```
python runuai.py --run=123
```
## Contact
For questions, contact us at:
andyshih at cs dot stanford dot edu