BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Interpretable representation learning for speech and audio signals
  - Dr Purvi Agrawal\, Indian Institute of Science (IISc) and Microsoft Ind
 ia
DTSTART:20210302T120000Z
DTEND:20210302T130000Z
UID:TALK157528@talks.cam.ac.uk
CONTACT:Dr Kate Knill
DESCRIPTION:The learning of interpretable representations from raw data pr
 esents significant challenges for time series data like speech. In this ta
 lk\, we will discuss a relevance weighting scheme that allows the interpre
 tation of the speech representations during the forward propagation of the
  model itself. \n* The relevance weighting is achieved in a 2-stage deep r
 epresentation learning framework where the weighting approach performs the
  task of feature selection at each stage. \n* A relevance sub-network\, ap
 plied on the first stage operating on raw speech signals\, acts as an acou
 stic filterbank layer with relevance weighting. A similar relevance sub-ne
 twork applied on the second convolutional layer performs modulation filter
 bank learning with relevance weighting.\n* All the layers are trained join
 tly for a speech recognition task on noisy and reverberant speech. The pro
 posed representation learning framework is also extended for the task of s
 ound classification. \n\nWe will discuss the detailed analysis of the rele
 vance weights and intermediate representations learned by the model which 
 would reveal that the relevance weights capture information regarding the 
 underlying speech/audio content\, along with improved system performances.
 \n\n*Bio:* Purvi Agrawal recently defended her Ph.D. thesis titled "Neural
  Representation learning for Speech and Audio Signals" from Learning and E
 xtraction of Acoustic Patterns (LEAP) lab with Dr. Sriram Ganapathy\, Dept
 . of Electrical Engineering\, Indian Institute of Science (IISc)\, Bangalo
 re. Prior to joining IISc\, she obtained her Masters in Speech Communicati
 ons from DA-IICT\, Gandhinagar in 2015. She has also worked in Sony R & D 
 Labs\, Tokyo in 2017. She will be joining as an Applied Researcher-II at M
 icrosoft India with the speech research team in Feb. 2021. Her research in
 terests include interpretable deep learning\, raw waveform modeling\, low-
 resource data modeling\, unsupervised/self-supervised learning.\n
LOCATION:Zoom: https://zoom.us/j/95352633552?pwd=RzJVK2UzOGZyNU5mVHd1Y1VPT
 2tDUT09
END:VEVENT
END:VCALENDAR