https://github.com/mitmul/chainer-nri
Reproduction work of "Neural Relational Inference for Interacting Systems" in Chainer
https://github.com/mitmul/chainer-nri
chainer deep-learning graph-neural-networks
Last synced: about 1 year ago
JSON representation
Reproduction work of "Neural Relational Inference for Interacting Systems" in Chainer
- Host: GitHub
- URL: https://github.com/mitmul/chainer-nri
- Owner: mitmul
- License: mit
- Created: 2019-01-23T00:20:56.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2019-02-05T03:46:55.000Z (over 7 years ago)
- Last Synced: 2025-03-18T21:54:31.321Z (about 1 year ago)
- Topics: chainer, deep-learning, graph-neural-networks
- Language: Python
- Size: 1.06 MB
- Stars: 32
- Watchers: 3
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural Relational Inference (NRI)
**Graph Neural Network for interacting systems**
Given a time series data of nodes, the NRI model predicts the future node states and underlying relashionship between the nodes as edges.
This is a reproduction work of the neural relational inference (NRI) in Chainer. The original implementation by the authors is found here: [ethanfetaya/NRI](https://github.com/ethanfetaya/NRI).
Please refer for details to the paper:
**Neural relational inference for interacting systems.**
Thomas Kipf*, Ethan Fetaya*, Kuan-Chieh Wang, Max Welling, Richard Zemel.
https://arxiv.org/abs/1802.04687 (*: equal contribution)
## Dataset
### Particle Physics Simulation Dataset
```bash
cd data
python generate_dataset.py
```
## Training
### Particle Physics Simulation Dataset
```bash
python train.py --gpu 0
```
## Visualize results
```bash
python utils/visualize_results.py \
--args-file results/2019-01-22_10-20-25_0/args.json \
--encoder-snapshot results/2019-01-22_10-20-25_0/encoder_epoch-500.npz \
--decoder-snapshot results/2019-01-22_10-20-25_0/decoder_epoch-500.npz \
--gpu 0
```
## Quantitative evaluation
### Accuracy (in %) of unsupervised interaction recovery
| Model | Springs - 5 nodes (test) |
|:---------------------------------------|:-------------------------|
| chainer-nri (MLPEncoder, MLPDecoder) | 99.8 |
| chainer-nri (CNNEncoder, MLPDecoder) | 99.4 |
| Original (from paper) | 99.9 |
### Mean squared error (MSE) in predicting future states for simulations with 5 nodes
| Model | Springs - 5 nodes (test) |
|:---------------------------------------|:-------------------------|
| chainer-nri (MLPEncoder, MLPDecoder) | 3.75e-05 |
| chainer-nri (CNNEncoder, MLPDecoder) | 3.83e-05 |
| Original (from paper) | 3.12e-08 |