We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.SD

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Sound

Title: Self-supervised Pre-training Reduces Label Permutation Instability of Speech Separation

Abstract: Speech separation has been well-developed while there are still problems waiting to be solved. The main problem we focus on in this paper is the frequent label permutation switching of permutation invariant training (PIT). For N-speaker separation, there would be N! possible label permutations. How to stably select correct label permutations is a long-standing problem. In this paper, we utilize self-supervised pre-training to stabilize the label permutations. Among several types of self-supervised tasks, speech enhancement based pre-training tasks show significant effectiveness in our experiments. When using off-the-shelf pre-trained models, training duration could be shortened to one-third to two-thirds. Furthermore, even taking pre-training time into account, the entire training process could still be shorter without a performance drop when using a larger batch size.
Comments: submitted to ICASSP2021
Subjects: Sound (cs.SD); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
Cite as: arXiv:2010.15366 [cs.SD]
  (or arXiv:2010.15366v1 [cs.SD] for this version)

Submission history

From: Sung-Feng Huang [view email]
[v1] Thu, 29 Oct 2020 06:07:01 GMT (1105kb,D)

Link back to: arXiv, form interface, contact.