2:30pm - 3:30pm

Thursday 13 October 2022

Learning with Little Data by Model Adaptation: Applications in Music, Sensors and Generative ML

CVSSP External Seminar by Dr. Vipul Arora, IIT Kanpur, India

Free

35BA00
CVSSP Seminar room
University of Surrey
Guildford
Surrey
GU2 7XH
back to all events

This event has passed

Speakers

  • Dr Vipul Arora

Learning with Little Data by Model Adaptation: Applications in Music, Sensors and Generative ML

Abstract:  

Deep-learning-based models achieve remarkable performances with labelled data. However, many practical scenarios face a scarcity of labelled data, while there may still be an abundance of unlabelled data. This talk will discuss several methods to learn effectively from limited data. These methods mostly fall under the paradigm of model adaptation and include fine-tuning-based transfer learning, meta-learning, and semi-supervised domain adaptation. These methods’ application to music melody estimation and sensor calibration (regression) problems will be demonstrated. Another way to learn from limited data is by using conditional models. This method will be illustrated for generative machine learning applied to XY models in statistical Physics and field theories in Particle physics. Time permitting, there will be a brief presentation on audio retrieval.

Short bio:

Vipul Arora is an Associate Professor at the Department of Electrical Engineering, IIT Kanpur, India. He received his B. Tech. and PhD degrees in Electrical Engineering from IIT Kanpur. He did a postdoc at Oxford University (UK), where I developed speech recognition systems using linguistic principles. Then he worked at Amazon in Boston (USA), where he worked on audio classification for developing the Alexa home security system. His current research interest is in developing machine learning algorithms for audio and signal processing. He works on model adaptation, uncertainty modelling and generative machine learning. http://home.iitk.ac.in/~vipular/

Attend the Event

This is a free online event to everyone. You can attend via Zoom

Visitor information


Hybrid event