[Colloquium] 11/9 TTIC Colloquium: Tara Sainath, Google Research

[Mary Marre]

When:     Monday, November 9th at 11:00 a.m.

Where:    TTIC, 6045 S Kenwood Avenue, 5th Floor, Room 526

Speaker:  Tara Sainath; Google Research

Title: Single and Multichannel Raw Waveform Neural Network Acoustic Models

Abstract:

Instead of starting with standard front-end features such as log-mel,
several
recent speech recognition results have demonstrated the possibility of
training neural network acoustic models directly on the time-domain
waveform.  Through supervised training, such networks are able to
learn a suitable auditory filterbank-like feature representation
simultaneously with a discriminative classifier, thereby eliminating
the need for hand crafted feature extraction.

In this talk I discuss research at Google towards this effort.
We have found that integrating using a CLDNN architecture with raw-waveform
modeling is critical,
and leads to a 3% relative reduction in word error
rate (WER) on noisy data compared to an analogous system trained on
mel features. Furthermore, similar waveform acoustic models trained
on multichannel waveforms can learn to do spatial filtering and be
robust to varying direction of arrival of the target speech signal.
Training such a network on inputs captured using multiple microphone
array configurations results in a system that is robust to a range of
microphone spacings, leading to a relative decrease of 11% WER
compared to a single channel system on data with mismatched spacing.

Host: Karen Livescu


For more information on the colloquium series or to subscribe to the
mailing list, please see http://www.ttic.edu/colloquium.php



Mary C. Marre
Administrative Assistant
*Toyota Technological Institute*
*6045 S. Kenwood Avenue*
*Room 504*
*Chicago, IL  60637*
*p:(773) 834-1757 <%28773%29%20834-1757>*
*f: (773) 357-6970 <%28773%29%20357-6970>*
*mmarre at ttic.edu <mmarre at ttic.edu>*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20151102/91921c24/attachment.htm