Internal code for dataset analysis: under construction.
See requirements.txt
Download the pre-trained BERT model from
here and unzip them inside the
pretrained_bert directory. In the paper, we use the cased base model.
To be added.
See each script for more documentation and descriptive command line flags.
python3 -m dataset_analysis.analyze_data.py: get high-level statistics of the data and correlation among targets.python3 -m dataset_analysis.extract_words.py: get the words that are significantly associated with each target, in contrast to the other targets, based on their log odds ratio.python3 -m dataset_analysis.ppca.py: run PPCA (Cowen et al., 2019)[https://www.nature.com/articles/s41562-019-0533-6] on the data and generate plots.
Run python -m dataset_analysis.bert_classifier.py to perform fine-tuning on top of
BERT, with added regularization. See the script and the paper for detailed
description of the flags and parameters.
If you use this code for your publication, please cite the original paper:
TODO: add citation