University of Cambridge > Talks.cam > Machine Learning Reading Group @ CUED > Deep Belief Networks for Phone Recongition

  Browse A-Z

Log in

Googledetails Microsoftdetails

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹ️ Information

Deep Belief Networks for Phone Recongition

Download to your calendar using vCal

If you have a question about this talk, please contact Shakir Mohamed .

The current state-of-the-art for acoustic models are Discriminatively trained Hidden Markov Models. There are proposals to use different types of model to improve upon the current state-of-the-art. One model is the Deep Belief Network that can produce a rich distributed representation of speech data. We describe Restricted Boltzmann Machines, how they are composed into a Deep Belief Network, and the application of a Deep Belief Network to phone recognition. If we have time we will touch on another deep structured acoustic model, the deep hidden conditional random field.

This is outlined in the paper http://www.cs.toronto.edu/~gdahl/papers/dbnPhoneRec.pdf

This talk is part of the Machine Learning Reading Group @ CUED series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Goldsmiths 1 & https://us06web.zoom.us/j/81688365420?pwd=b3RpdHgxNVhWbWQrbXluMUxYNWVtQT09 Meeting ID: 816 8836 5420 Passcode: 047859 In person in the Max Perutz Lecture Theatre (CB2 0QH) and via Zoom, link:https://mrc-lmb-cam-ac-uk.zoom.us/j/97590440299?pwd=d3k2U3l1WENLbllGT1ZUODYyd3lFZz09 Henry Wellcome Building, Fitzwilliam Street, Cambridge

Other talks

CPH IRC and PHI Title TBC César Milstein Lecture - Title TBC Disentangling preventative & ejective AGN feedback on a population basis with simulations, observations & machine learning title and abstract tba (joint seminar with Spectral Geometry in the Clouds) Topology of the Ritz energy landscape
 

© 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity