We are glad to have Charbel Sakr as social hour guest speaker this Friday, April 13th at 3:00PM in 369 CSL, In this week’s social hour Charbel will talk about “Reduced Precision Machine Learning“.
Abstract: Machine learning models such as deep neural networks are starting to achieve human level accuracy in many cognitive tasks. However, such successes often overshadow the tremendous complexity of typical deep learning models. In this talk, we focus on numerical precision – a key parameter defining the complexity of neural networks.
Bio: Charbel Sakr is a graduate student working in CSL with Professor Naresh Shanbhag. Charbel obtained his Engineering degree from the American University of Beirut in 2015 graduating with High Distinction. He then joined the University of Illinois where he is now a PhD candidate in the department of Electrical and Computer Engineering. His research interests are in resource-constrained machine learning, with a focus on analysis and implementation of reduced precision algorithms and models. Charbel is also interested in fun facts and coincidences. For instance, he takes great joy in knowing that, should he ever have his own lab, its acronym would be CSL.