Findings from ieeexplore.ieee.org/document/106... show that speech SSL models converge on similar embedding spaces, but via different routes. While overall representations align, individual neurons learn distinct localized concepts.
Interesting read! @fbk-mt.bsky.social
ieeexplore.ieee.org
Speech Representation Analysis Based on Inter- and Intra-Model Similarities
Self-supervised models have revolutionized speech processing, achieving new levels of performance in a wide variety of tasks with limited resources. However, the inner workings of these models are sti...
