Biologically-Inspired Unsupervised Learning of Higher Order Visual Features

News

With invitation of Computer Science and Software Engineering, Prof Marwan Jabri provided a seminar on Monday afternoon, 20 April 2015, with a topic "Biologically-Inspired Unsupervised Learning of Higher Order Visual Features".

Professor Marwan Jabri has developed an architecture and unsupervised learning algorithms for visual pattern recognition, inspired from the primate visual neural processing pathway. The architecture and the algorithm is an important development to understand how higher order visual feature detectors (tuned processing elements) develop. With task of face recognition from the LFW and proprietary datasets, the model demonstrates very encouraging results on fairly challenging recognition conditions, which include multiple facial poses, illuminations/brightness, and face rotations, with over 89% success rate on LFW datasets and 95% on a proprietary dataset.

Professor Jabri is currently Founder and Chief Executive at Neuromorphic Systems LLC, an early stage company developing intelligent agents. Prof. Jabri brings nearly twenty-five years of research and development in neural computation, machine learning, multimedia communications systems, and audio/video compression and transcoding. Dr. Jabri was the founder of Dilithium Networks Inc. which developed and deployed mobile video solutions to over 100 global mobile operators and service providers worldwide. Dilithium was acquired in 2010 by OnMobile Global Ltd, a leading provider of mobile value added services.

Prior to Neuromorphic and Dilithium, Dr Jabri was the Gordon and Betty Moore Endowed Chair Professor at OGI/OHSU, Portland Oregon, and Professor of Adaptive Systems, University of Sydney, Australia. As academic, Dr Jabri researched and developed neural computing architectures, learning algorithms, and integrated systems for pattern recognition in the areas of biomedical engineering, robotics, financial instrument forecasting, and still- and moving-picture recognition.

Professor Jabri was also awarded the Outstanding Young Investigator Medal by the Australian Telecommunication and Electronics Research Board for contributions to the application of artificial intelligence techniques to real-world problems, Innovation Technology Award by the World Economic Forum, and is member of the Experts Network of the Australian Federal Government Commercialisation Australia program. As CTO of Dilithium networks, Dr Jabri received multiple awards for Dilithium products.

Professor Jabri holds a License de Physique and a Maîtrise de Physique, University of Paris Diderot, France, and a Ph.D. in Electrical Engineering from the University of Sydney. Dr. Jabri has been a Visiting Scientist at the Salk Institute, San Diego, California and at Bell Laboratories, New Jersey. Dr. Jabri is the author or co-author of 2 books, 41 granted patents, and over 150 technical publications in leading international journals and conferences.

Tagged as: Research

Share