ICML2018 - Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction

The algorithm is described in a ICML 2018 paper (supplementary materials).

The generalized Earley parser is implemented in src/python/parser/generalizedearley.py.

Folders in the repository:

src/python/datasets: code for loading the extracted features of the used datasets.
src/python/experiments: code for running the experiments. The main experiments are implemented in cad.py and wnp.py.
src/python/parser: files for grammar inference.

If you find this code useful, please cite our work with the following bibtex:

@inproceedings{qi2018future,
    title={Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction},
    author={Qi, Siyuan and Jia, Baoxiong and Zhu, Song-Chun},
    booktitle={International Conference on Machine Learning (ICML)},
    year={2018}
}