Inducing brain-relevant bias in natural language processing models (1911.03268v1)

Abstract: Progress in NLP models that estimate representations of word sequences has recently been leveraged to improve the understanding of language processing in the brain. However, these models have not been specifically designed to capture the way the brain represents language meaning. We hypothesize that fine-tuning these models to predict recordings of brain activity of people reading text will lead to representations that encode more brain-activity-relevant language information. We demonstrate that a version of BERT, a recently introduced and powerful LLM, can improve the prediction of brain activity after fine-tuning. We show that the relationship between language and brain activity learned by BERT during this fine-tuning transfers across multiple participants. We also show that, for some participants, the fine-tuned representations learned from both magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI) are better for predicting fMRI than the representations learned from fMRI alone, indicating that the learned representations capture brain-activity-relevant information that is not simply an artifact of the modality. While changes to language representations help the model predict brain activity, they also do not harm the model's ability to perform downstream NLP tasks. Our findings are notable for research on language understanding in the brain.