COLA: Contextualized Commonsense Causal Reasoning from the Causal Inference Perspective

This repository is the official implementation of COLA: Contextualized Commonsense Causal Reasoning from the Causal Inference Perspective.

The paper is accepted to the main conference of ACL 2023.

<div align="center"> <img src="method_figure.jpg" alt="Model"/> <span>An Overview of Our Method</span> </div>

Requirements

Python version is 3.8.5

requirements:

allennlp_models==2.10.1
datasets==2.6.1
evaluate==0.3.0
matplotlib==3.6.1
numpy==1.23.4
pandas==1.5.0
pathos==0.2.9
statsmodels==0.13.2
torch==1.10.2
tqdm==4.64.1
transformers==4.20.0
ujson==5.5.0

You can install all requirements with the command

pip install -r requirements.txt

Datasets

Data for Temporal Predictor

We extract a large dataset containing 800K pairs of events and their temporal order from RocStories. You can download it here download All licenses are subject to RocStories' original release.

Choice of Plausible Event in Sequence (COPES)

The contextualized commonsense causal reasoing dataset COPES is in the dir ./COPES_data.

Run COLA Framework

Our COLA Framework owns a few steps to conduct inference.

Step1: Fine-tune a Temporal Predictor with incidental supervision signals

First, you need to fine-tune a temporal predictor, which learns temporal relations. The code for this part is in mlm.py and mlm_util.py.

We provide an example command in the script run_script/finetune_mlm.sh. It fine-tunes bert-base-uncased, you can change it to other models.

Step2: Sample Covariates

Second, you need to sample covariates for each event. Run the script run_script/sample_covariates.sh.

After getting the covariates, you need to do Multistamp Covariate Sampling, which tasks the union of covariates sampled at multiple timestamps. Run the example command

python mix_covariate.py --input_path YOUR_COVARIATES --output_path YOUR_OUTPUT_PATH --copes_path ./COPES_data/COPES.json

Step3: Sample Interventions

To generate interventions of events, you first need to run SRL tool to extract verbs and two arguments ARG0 and ARG1:

python SRL_parsing.py --data_path ./COPES_data/COPES.json \
--model_path structured-prediction-srl-bert --batch_size 512 \
--output_dir YOUR_OUTPUT_DIR

Step4: Temporal Relation Reasoning

After generating covariates and interventions, you can predict the temporal relations between covariates, interventions and RocStories events.

For temporal relations in covariate balancing (between covariates and interventions/original causes), run run_script/get_cov-inter.sh.

For average treatment effect (between interventions/original causes and caused events), run run_script/get_inter-outcome.sh.

Conduct Inference

Finally, you can conduct causal inference based on temporal relations: run run_script/evaluate.sh.

Pre-Computed Results

Some operations are computationally heavy (e.g., using GPT-J-6b), you can download our pre-computed results here download Temporal relations are predicted by fine-tuned bert-base-uncased.

Contributing

If you find any typo or bug, please open an issue. This repo is maintained by Zhaowei Wang