Opiate Addiction Identification (OpiateID)

Codebase for the NAACL 2024 Findings paper "Identifying Self-Disclosures of Use, Misuse and Addiction in Community-based Social Media Posts"

Main Authors:

1. Chenghao Yang (yangalan1996@gmail.com) (University of Chicago, Previously at Columbia University)

2. Tuhin Chakrabarty (tuhin.chakr@cs.columbia.edu) (Columbia University)

Supervisor Team:

1. Nabila El-Bassel (School of Social Work, Columbia University)

2. Smaranda Muresan (Data Science Institute, Columbia University)

Reference

If you use this code as part of any published research, please acknowledge the following paper (it encourages researchers who publish their code!):

@inproceedings{yang-2023-identifying,
    title = "Identifying Self-Disclosures of Use, Misuse and Addiction in Community-based Social Media Posts",
    author = "Chenghao Yang and Tuhin Chakrabarty and Karli R Hochstatter and Melissa N Slavin and Nabila El-Bassel and Smaranda Muresan",
    booktitle = "Findings of the Association for Computational Linguistics: NAACL 2024",
    year = "2024",
    publisher = "Association for Computational Linguistics",
}

Project Structure

1. codebase: Codebase for Data Processing, Fine-tuning, evaluation and visualization.
2. data: The data for prompting and evaluation. As per IRB ethics approval, we kindly request the user to submit a request here to explain the project scope and obtained ethics approval before we send you the access to data.

Dependency Installation

For the main repo:

conda create -p ./env python=3.9
conda activate ./env # the environment position is optional, you can choose whatever places you like to save dependencies. Here I choose ./env for example.
pip install -r requirements.txt

Running Instructions

Section 4: Main Experiments

1. Prompting GPT3.5/4: check out codes in codebase/prompt_gpt4.py for running prompts and collecting responses. Then run post_processing_log.py to do necessary postprocessing for normalizing the model outputs.
2. Fine-tuning DeBERTa: check out codes in codebase/deberta_finetuning.py.

Section 5: The Role of Explanations

Check out codes in codebase/create_sliver_rational_annotation_files.py and codebase/process_finetune_data_for_finetuning.py to see how to combine gold rationale, sliver rationale and random rationale. Here we would re-use the codes in Section 4 for evaluation.

Section 6: Error Analysis and the Influence of Dataset Annotation Uncertainty

Check out codes in codebase/error_analysis.py.