Anonymized BERT: An Augmentation Approach to the Gendered Pronoun Resolution Challenge
This addresses pronoun resolution in NLP, offering a practical solution to reduce gender and regional biases in names, though it is incremental as it builds on existing BERT without fine-tuning.
The paper tackled the Gendered Pronoun Resolution challenge by developing an augmentation method that anonymizes candidate names with common placeholders, achieving a log loss of 0.1947 (7th place) with a 0.04 improvement from augmentation, and post-competition analysis showed potential for 0.1799 (3rd place).
We present our 7th place solution to the Gendered Pronoun Resolution challenge, which uses BERT without fine-tuning and a novel augmentation strategy designed for contextual embedding token-level tasks. Our method anonymizes the referent by replacing candidate names with a set of common placeholder names. Besides the usual benefits of effectively increasing training data size, this approach diversifies idiosyncratic information embedded in names. Using same set of common first names can also help the model recognize names better, shorten token length, and remove gender and regional biases associated with names. The system scored 0.1947 log loss in stage 2, where the augmentation contributed to an improvements of 0.04. Post-competition analysis shows that, when using different embedding layers, the system scores 0.1799 which would be third place.