Untangling in Invariant Speech Recognition
Name
NeurIPS-2019-untangling-in-invariant-speech-recognition-Paper.pdf
Description
Published version
Size
1.97 MB
Format
Adobe PDF
Checksum (MD5)
b06b6bcea8ee47b0f7f95f5283cbdd37
Author(s)
Stephenson, Cory
Feather, Jenelle
Padhy, Suchismita
Elibol, Oguz
Tang, Hanlin
McDermott, Josh
Chung, SueYeon
Date Issued
2019
Journal
Advances in Neural Information Processing Systems
Citation
Stephenson, Cory, Feather, Jenelle, Padhy, Suchismita, Elibol, Oguz, Tang, Hanlin et al. 2019. "Untangling in Invariant Speech Recognition." Advances in Neural Information Processing Systems, 32.
Version
Final published version
Abstract
© 2019 Neural information processing systems foundation. All rights reserved. Encouraged by the success of deep neural networks on a variety of visual tasks, much theoretical and experimental work has been aimed at understanding and interpreting how vision networks operate. Meanwhile, deep neural networks have also achieved impressive performance in audio processing applications, both as sub-components of larger systems and as complete end-to-end systems by themselves. Despite their empirical successes, comparatively little is understood about how these audio models accomplish these tasks. In this work, we employ a recently developed statistical mechanical theory that connects geometric properties of network representations and the separability of classes to probe how information is untangled within neural networks trained to recognize speech. We observe that speaker-specific nuisance variations are discarded by the network's hierarchy, whereas task-relevant properties such as words and phonemes are untangled in later layers. Higher level concepts such as parts-of-speech and context dependence also emerge in the later layers of the network. Finally, we find that the deep representations carry out significant temporal untangling by efficiently extracting task-relevant features at each time step of the computation. Taken together, these findings shed light on how deep auditory models process time dependent input signals to achieve invariant speech recognition, and show how different concepts emerge through the layers of the network.
MIT Department
Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences
Center for Brains, Minds, and Machines
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://papers.nips.cc/paper/2019/hash/e2db7186375992e729165726762cb4c1-Abstract.html
