Hi All,
This week we will have Navdeep Jaitly, a student with Geoffrey Hinton, who will tell us about his work on Transforming Autoencoders. See you there.
Where: LISA lab (AA3256) When: 14h00, Thursday, July 28th 2011
Title: Transforming Autoencoders for speech.
Speech recognition is dominated by use of Gaussian Mixture Model / Hidden Markov Models trained on mel frequency cepstral representations of speech. Psychologists have long argued that highy-quality recognition would be facilitated by finding acoustic events or landmarks that have well-defined onset times, amplitudes and rates in addition to being present or absent. We introduce a new way of learning such acoustic events by using a new type of autoencoder that is given both a spectrogram and desired global transformation and learns to output the transformed spectrogram. By specifying the global transformation in the appropriate way, we can force the autoencoder to extract accoustic events that, in addition to a probability of being present, have explicit onset times, amplitudes and rates. This makes it much easier to compute relationships between acoustic events. I'll talk about the work we have done and the directions we are currently pursuing.
This is joint work with Geoffrey Hinton.
Cheers, Aaron
Afficher les réponses par date