Learning and Optimization Problems in Audio Personalization

Thumbnail

Event details

Date 06.08.2024
Hour 11:0012:00
Speaker Professor Romit Roy Choudhury, University of Illinois at Urbana Champaign (UIUC)
Location
Category Conferences - Seminars
Event Language English
Abstract: We will discuss problems in audio personalization for humans, in the context of earphones, hearing aids, or glasses. Many systems or models are trained based on public datasets, however, personalizing those models to a specific user often brings the burden of collecting that user’s data. How can we reduce this burden? We will discuss some of our recent work on (1) self-supervised audio denoising for earphones, and (2) black-box optimization methods for personalizing hearing aids (or any other content for that matter) from a limited number of user’s feedback. We will spend the last part of the talk on potentially generalizing the ideas to other applications that audience members may be working on.

Bio: Romit Roy Choudhury is a Gilmore Family Endowed Professor of ECE and CS at the University of Illinois at Urbana Champaign (UIUC). He joined UIUC from Fall 2013, prior to which he was an Associate Professor at Duke University. Romit received his PhD in the CS department of UIUC in Fall 2006. His current research interests are in generative models with applications to audio and RF sensing. Along with his students, he received a few research awards, including the ACM Sigmobile Rockstar Award, the UIUC Distinguished Alumni Award, the Google Research Award, etc. He was elevated to IEEE Fellow in 2018. He is also an Amazon Scholar since 2022.

Practical information

  • Informed public
  • Free

Organizer

  • Professor Haitham Al Hassanieh EPFL IC IINFCOM SENS

Event broadcasted in

Share