IEEE-CASS DLP Lecture at IMSE-CNM (CSIC/University of Seville)

Last September 16, 2016, Prof. Shih-Chii Liu, Distinguished Lecturer of the IEEE Circuits and Systems Society (CASS), visited the Institute of Microelectronics of Seville, IMSE-CNM (CSIC/University of Seville), in Seville, Spain, where she gave a talk entitled “Event-Based Auditory Processing with Spiking Silicon Cochleas and Deep Networks”.

This event was sponsored by the IEEE-CASS Distinguished Lecturer Program, and it was locally organized by the Spain Chapter of IEEE-CASS, chaired by Prof. Jose M. de la Rosa. The event included also a visit to the IMSE-CNM labs.

 

See some pictures of the event at: 
http://www2.imse-cnm.csic.es/~jrosa/DLP_Liu_IEEE-CASS_SpainChapter_News_160916.pdf

News published at IEEE-CASS Newsletter, vol. 10, issue 5:

http://cassnewsletter.org/Volume10-Issue5/society-news.html​

 

LECTURE DETAILS

 

Title 

 

Event-Based Auditory Processing with Spiking Silicon Cochleas and Deep Networks

 

Abstract

 

Audio processing based on conventional regular sampling, process audioframes unnecessarily even when the frames carry no information. They also require high sampling rates for auditory scene parsing where source localization and separation are essential. Event-based neuromorphic audio sensors and processing algorithms offer a potential solution to these applications for IoT, mobile, and always-on applications by asynchronously sampling and processing the audio input in a data driven way. This talk covers the latest audio sensing systems including a new sub milliwatt binaural silicon cochlea, event-based algorithms that process the outputs of these cochlea sensors, and example system applications such as auditory localization using a factor of 40 less computing power than conventional Nyquist-rate systems. The talk also covers event-driven deep networks that use the output of the cochleas and the impact of bit precision of such networks on their performance.

 

Lecturer Biography

 

Dr. Shih-Chii Liu co-leads the Sensors Group at the Institute of Neuroinformatics, University of Zurich ant ETH Zurich, Switzerland. She studied electrical engineering as an undergraduate and received the PhD degree in the Computation and Neural Systems program from the California Institute of Technology. She worked at various companies including Gould American Microsystems, LSI Logic, and Rockwell International Research Labs. Her group has been working on the design of low-power neuromorphic event-based auditory and visual sensors, and research into neuromorphic algorithms and event-driven machine learning deep networks for processing inputs from these sensors. These algorithms and networks have been implemented on event-driven hardware systems that operate in natural environments.

Dr. Liu is past Chair of the IEEE CAS Sensory Systems and Neural Systems and Applications Technical Committees. She is current Chair of the IEEE Swiss CAS/ED Society and an Associate Editor of the IEEE Transactions of Biomedical Circuits and Systems and Neural Networks journal.