README.md

This is the repo of code and data for the EMNLP 2021 paper "The Causal Direction of Data Collection Matters: Implications of Causal and Anticausal Learning for NLP" by Zhijing Jin*, Julius von Kügelgen*, Jingwei Ni, Tejas Vaidhya, Ayush Kaushal, Mrinmaya Sachan, Bernhard Schoelkopf.

To cite the paper, use the BibTex below:

@inproceedings{jin2021causal,
  author    = {Zhijing Jin
and Julius von Kuegelgen
and Jingwei Ni
and Tejas Vaidhya
and Ayush Kaushal
and Mrinmaya Sachan
and Bernhard Schoelkopf},
  title     = {Causal Direction of Data Collection Matters: {I}mplications of Causal and Anticausal Learning for {NLP}},
  booktitle = {Proceedings of the 2021 Conference on Empirical Methods in Natural
               Language Processing, {EMNLP} 2021},
  publisher = {Association for Computational Linguistics},
  year      = {2021},
  url       = {https://arxiv.org/abs/2110.03618},
}

Structure

• Code for MDL (paper Section 4.2): mdl/ with data in this zip file (24.3MB)
• Code for deciperment experiments (paper Section 5.1): decipher/
• Code for meta study significance tests (paper Section 5.2 & 6): meta_study_significance_test.py with annotated data in this Google Spreadsheet