2024/06/12 by Hengyu Li, Kangdi Mei, Li, Hengyu +11 · 1 citation
Computer Science · Neuroscience · #Audio and Speech Processing (eess.AS) #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #FOS: Electrical engineering #Intelligent Tutoring Systems and Adaptive Learning #Sound (cs.SD) #Speech and dialogue systems #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2406.08266
openalex publication_date 2024/06/12 · openalex created_date 2024/06/15 · openalex updated_date 2026/07/28
It was shown in literature that speech representations extracted by self-supervised pre-trained models exhibit similarities with brain activations of human for speech perception and fine-tuning speech representation models on downstream tasks can further improve the similarity. However, it still remains unclear if this similarity can be used to optimize the pre-trained speech models. In this work, we therefore propose to use the brain activations recorded by fMRI to refine the often-used wav2vec2.0 model by aligning model representations toward human neural responses. Experimental results on SUPERB reveal that this operation is beneficial for several downstream tasks, e.g., speaker verification, automatic speech recognition, intent classification.One can then consider the proposed method as a new alternative to improve self-supervised speech models.