CausalChaos! Dataset for Video Question Answering - [NeurIPS 2024]

CausalChaos! Dataset for Comprehensive Causal Action Question Answering Over Longer Causal Chains Grounded in Dynamic Visual Scenes

Contents

We provide the following:

1. Dataset (Annotations, Precomputed Visual Features, Synchronization frame numbers)
2. Code

Feel free to request any particular type of features you might be interested in by emailing us or opening issue/pull request!

Citation

Please consider citing our work:

@article{causalchaos2024,
  title={CausalChaos! Dataset for Comprehensive Causal Action Question Answering Over Longer Causal Chains Grounded in Dynamic Visual Scenes},
  author={Parmar, Paritosh and Peh, Eric and Chen, Ruirui and Lam, Ting En and Chen, Yuhan and Tan, Elston and Fernando, Basura},
  journal={arXiv preprint arXiv:2404.01299},
  year={2024}
}