Generating Hypothetical Events for Abductive Inference

This directory contains the following parts of the 'Generating Hypothetical Events for Abductive Inference' experiment.

Along with our code we include the relevant datasets used in the paper. The TIMETRAVEL dataset are taken from website: https://drive.google.com/file/d/150jP5FEHqJD3TmTO_8VGdgqBftTDKn4w/view

We train GPT-2 model to generate a possible event that could happen next, given some counterfactual scenarios for a given story. We used the following script to prepare our training : create_counterfactual_data.py

Preprocessed data can be found in data/counterfactual

Unsupervised Setting

We used BERT score to evaluate our hypothesis that the generated possible next event observation (O^{2}{H{j}}) given the more plausible hypothesis H_{j} should be more similar to observation {O_2}.

pip install bert-score

Run the following script to get the bert score on anli test:

get_bert_score.py output_file data/anli/test.jsonl data/anli/test-labels.lst

Citation

@inproceedings{paul-frank-2021-generating,
    title = "Generating Hypothetical Events for Abductive Inference",
    author = "Paul, Debjit  and
      Frank, Anette",
    booktitle = "Proceedings of *SEM 2021: The Tenth Joint Conference on Lexical and Computational Semantics",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.starsem-1.6",
    doi = "10.18653/v1/2021.starsem-1.6",
    pages = "67--77"
    }