# Gendered Pronoun Resolution using BERT and an extractive question answering formulation

## TL;DR

This paper introduces a BERT-based extractive question answering approach for gendered pronoun resolution, significantly reducing gender bias and outperforming existing methods without hand-engineered features.

## Contribution

It proposes a novel QA formulation for pronoun resolution that minimizes gender bias and enhances performance using fine-tuned BERT models and ensemble techniques.

## Key findings

- Achieved 22.2% absolute improvement in F1 score over baseline.
- Reduced gender bias to 0.99 on the dataset.
- Ensemble model ranked 9th in ACL workshop shared task.

## Abstract

The resolution of ambiguous pronouns is a longstanding challenge in Natural Language Understanding. Recent studies have suggested gender bias among state-of-the-art coreference resolution systems. As an example, Google AI Language team recently released a gender-balanced dataset and showed that performance of these coreference resolvers is significantly limited on the dataset. In this paper, we propose an extractive question answering (QA) formulation of pronoun resolution task that overcomes this limitation and shows much lower gender bias (0.99) on their dataset. This system uses fine-tuned representations from the pre-trained BERT model and outperforms the existing baseline by a significant margin (22.2% absolute improvement in F1 score) without using any hand-engineered features. This QA framework is equally performant even without the knowledge of the candidate antecedents of the pronoun. An ensemble of QA and BERT-based multiple choice and sequence classification models further improves the F1 (23.3% absolute improvement upon the baseline). This ensemble model was submitted to the shared task for the 1st ACL workshop on Gender Bias for Natural Language Processing. It ranked 9th on the final official leaderboard. Source code is available at https://github.com/rakeshchada/corefqa