An Introduction to Neuromorphic Computing with Intel Loihi

Speaker: Ashish Rao Mangalore, Doctoral resident at Intel Labs, Munich, Germany

Neuromorphic computing is a new paradigm of computing inspired by the organization and functioning of neurons in the mammalian brain. The architectural features derived from this inspiration result in orders of gain in the time and energy to solution for various classes of problems. Over the years, there have been neuromorphic platforms of different types, e.g., IBM TrueNorth, SpiNNaker, DYNAPs, and BrainChip Akida to name a few. In this talk, we shall focus on the state-of-the-art digital CMOS-based neuromorphic research platform, Loihi, developed by Intel Corporation. We shall then go through the basic operating principles of Loihi, problems that Loihi is best suited for, and the current main algorithmic research verticals being pursued by the neuromorphic research community. After this, there shall be a special emphasis on mathematical optimization on Loihi, one of the most promising and viable applications on Loihi. The talk will end with pointers on how groups can get involved to address open problems and contribute to neuromorphic research in general.

Biography of the speaker:

Ashish Rao Mangalore currently works as a doctoral resident at Intel Labs, Munich, Germany under the supervision of Prof. Alin Albu-Schäffer at the German Aerospace Center/DLR & TU Munich. His main research focus lies in the development & implementation of control algorithms for robotics and mathematical optimization problems on neuromorphic computers, specifically, Loihi. During a prior research stint at the Indian Institute of Science (IISc), Bengaluru, he conducted research on 3-D reconstruction techniques with neuromorphic cameras. He obtained his masters in Neuroengineering from the Technical University of Munich (TUM) and bachelors in Electrical & Electronics Engineering from R.V. College of Engineering (RVCE), Bengaluru.