The talk will discuss the challenging problem of designing Deep Neural Network systems that achieve high performance under low power envelopes, hindering their deployment in the embedded space. The talk will specifically focus on the opportunities that are provided when customisation of the design is possible, through the use of reconfigurable computing, and recent efforts on the challenges in automating the design of those systems will be discussed. Our recent work on the development of toolflows for automating the mapping of single and multiple CNNs into FPGAs will be discussed in detail, showcasing our efforts in addressing the above challenges.
Christos-Savvas Bouganis is a Reader in Intelligent Digital Systems in the Department of Electrical and Electronic Engineering, Imperial College London, U.K. He is leading the iDSL group at Imperial College, with a focus on the theory and practice of reconfigurable computing and design automation, mainly targeting the domains of Machine Learning, Computer Vision, and Robotics.