Auto-NPP

✨Automatic Integration for Neural Point Process✨

| Paper

Automatic Integration for Fast and Interpretable Neural Point Processes

| Installation

Dependencies: make, conda-lock

make create_environment
conda activate autonpp

| Dataset Download

make download prefix=data

| Get Trained Models

make download prefix=models

| Training and Testing

Specify the parameters in configs/test_autoint_1d_dataset.yaml and then run

make run

The loss curves and example intensity predictions are saved to figs/. With real-world datasets, the ground truth intensity is a placeholder and can be safely ignored. The logs are saved to logs/. The models are saved to models/.

To use the trained models, set retrain: false.

| Cite

@article{zhou2023automatic,
  title={Automatic Integration for Fast and Interpretable Neural Point Processes},
  author={Zhou, Zihao and Yu, Rose},
  journal={Learning for Dynamics and Control (L4DC)},
  year={2023}
}