Technical Talk on: Deep Belief Networks for Speaker Recognition

Share

Technical Talk on: Deep Belief Networks for Speaker Recognition



  Date and Time

  Location

  Hosts

  Registration



  • Date: 01 May 2015
  • Time: 11:00 AM to 12:00 PM
  • All times are (GMT+10:00) Australia/Queensland
  • Add_To_Calendar_icon Add Event to Calendar
  • Queensland University of Technology
  • 2, George Street, Brisbane
  • Brisbane, Queensland
  • Australia 4001
  • Building: S-Block
  • Room Number: 620
  • Click here for Map

  • Co-sponsored by IEEE Queensland Section


  Speakers

Omid Ghahabi

Topic:

Deep Belief Networks for Speaker Recognition

Speaker recognition is the process of automatically recognizing who is speaking by using the speakerspecific information included in speech signals. Applicable services for speaker recognition include voice dialing, banking over a telephone network, telephone shopping, database access services, information and reservation services, voice mail, security control for confidential information, remote access to computers as well as its application as a forensic tool. Deep Belief Networks (DBNs) have recently opened a new research line in image, audio, and speech processing areas. DBNs are originally generative network models which can be trained by a greedy layer-wise algorithm using Restricted Boltzmann Machines (RBMs). However, by adding a top label layer and using a standard backpropagation algorithm, these generative DBNs can be converted to discriminative ones what is often called a pre-trained discriminative network. Acoustic modeling using DBNs has been shown to be effective in speech recognition. However, few attempts have been carried out in speaker recognition area. The talk will first give a brief overview on speaker recognition and DBN. Then a summary of the PhD thesis carried out on the application of RBMs and DBNs for speaker recognition will be presented.


 

Biography:

Omid Ghahabi received the M.Sc. Degree in electrical engineering from Shahid Beheshti University, Tehran, Iran, in 2009. From 2009 to 2011, he has been with the speech processing group of the Research Center of Intelligent Signal Processing (RCISP), Tehran, Iran. He is now a Ph.D. candidate at Universitat Politecnica de Catalunya (UPC) - BarcelonaTech, Spain. He is working as a researcher in the speech processing group of the Signal Theory and Communications Department of UPC. He is also a member of the Research Center for Language and Speech Technologies and Applications (TALP), Barcelona, Spain. His research interests include speech processing, speaker recognition, and the applications of Deep Learning for speech and speaker recognition. He is the author of several journal and conference papers on these topics.

Email:

Address:Signal Theory and Communications Department, Universitat Politècnica de Catalunya (UPC),, Campus Nord, D5-120, Jordi Girona, 1-3, Barcelona, Spain, 08034

Omid Ghahabi

Topic:

Deep Belief Networks for Speaker Recognition

Biography:

Email:

Address:Barcelona, Spain