Gender-Neutral Large Language Models for Medical Applications: Reducing Bias in PubMed Abstracts
This addresses bias in medical AI applications, but it is incremental as it focuses on pronoun replacement in a specific dataset.
The paper tackled gender bias in large language models for medical literature by developing a pipeline to neutralize gendered occupational pronouns in PubMed abstracts, resulting in a model achieving a 70% inclusive replacement rate compared to 4% for a baseline.
This paper presents a pipeline for mitigating gender bias in large language models (LLMs) used in medical literature by neutralizing gendered occupational pronouns. A dataset of 379,000 PubMed abstracts from 1965-1980 was processed to identify and modify pronouns tied to professions. We developed a BERT-based model, "Modern Occupational Bias Elimination with Refined Training," or "MOBERT," trained on these neutralized abstracts, and compared its performance with "1965BERT," trained on the original dataset. MOBERT achieved a 70% inclusive replacement rate, while 1965BERT reached only 4%. A further analysis of MOBERT revealed that pronoun replacement accuracy correlated with the frequency of occupational terms in the training data. We propose expanding the dataset and refining the pipeline to improve performance and ensure more equitable language modeling in medical applications.