Noise-Contrastive Estimation for Multivariate Point Processes

Source code for Noise-Contrastive Estimation for Multivariate Point Processes (NeurIPS 2020).

Reference

If you use this code as part of any published research, please acknowledge the following paper (it encourages researchers who publish their code!):

@inproceedings{mei-2020-nce,
  author =      {Hongyuan Mei and Tom Wan and Jason Eisner},
  title =       {Noise-Contrastive Estimation for Multivariate Point Processes},
  booktitle =   {Advances in Neural Information Processing Systems},
  year =        {2020}
}

Instructions

Here are the instructions to use the code base.

Dependencies and Installation

This code is written in Python 3, and I recommend you to install:

• Anaconda that provides almost all the Python-related dependencies;

Run the command line below to install the package (add -e option if you need an editable installation):

pip install .

It will automatically install the following important dependencies:

• PyTorch 1.1.0 that handles auto-differentiation.

Prepare Data

The datasets used in our experiments can be downloaded from this Google Drive directory.

Place the datasets into:

data

Train Models

Go to

ncempp/run

To train models with MLE or NCE, try the command line below for detailed guide:

python train.py --help

The training logs and model parameters are stored in this directory:

logs/DATA_NAME/PARAMS

Draw Curves

To organize training logs and draw learning curves, use the command lines below for detailed guide:

python org_log.py --help

python draw_lc.py --help

License

This project is licensed under the MIT License - see the LICENSE file for details.