WildGraph: Realistic Graph-based Trajectory Generation for Wildlife

This repository contains the official implementation of WildGraph: Realistic Graph-based Trajectory Generation for Wildlife.

Train

Generate

Getting Started

1. Clone this repository:

git clone [https://github.com/aliwister/wildgraph.git](https://github.com/aliwister/wildgraph.git)
cd wildgraph

2. Create a conda environment and install the dependencies:

conda create --name wildgraph --file requirements.txt

Training

It is quite easy to train and test WildGraph or any of the benchmark methods reported:

python wild_run.py --dataset <geese|stork> --exp <WILDGRAPH|GAN|VAE|WILDGEN|TRANSFORMER> --epochs --split_distance --num_exps --desc --ablate<uniform_coarse|uniform_fine|no_pe|bow>

To train WildGraph:

python wild_run.py --dataset geese --exp WILDGRAPH --epochs 90 --split_distance .25

To train the VAE Benchmark:

python wild_run.py --dataset geese --exp VAE --epochs 90 

After training, a report will be saved in wild_experiments_log/[EXP] automatically.

Citation

If you found this repository useful, please consider citing our work:

@inproceedings{al2024wildgraph,
  title={WildGraph: Realistic Long-Horizon Trajectory Generation with Limited Sample Size},
  author={Al-Lawati, Ali and Eshra, Elsayed and Mitra, Prasenjit},
  booktitle={Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems},
  pages={247--258},
  year={2024}
}

License

This repository is licensed under Apache 2.0.