An efficient encoder-decoder architecture with top-down attention for speech separation.

Kai Li, Runxuan Yang, Xiaolin Hu

Browse the full ICLR paper archive.