TrueNorth Ecosystem for Brain-Inspired Computing: Scalable Systems, Software, and Applications
- Jun Sawada
- Filipp Akopyan
- et al.
- SC 2016
Filipp's main interests are in power-efficient VLSI designs that operate under extreme conditions and withstand process variations. Filipp is currently leading a team of highly-skilled engineers developing resilient synchronous-asynchronous chips, systems, tool flow, firmware for neural networks and signal processing.
Filipp Akopyan is a Principal Research Scientist, based at the IBM Thomas J. Watson Research Center. Currently he is the Chip Physical Design Technical Lead, as well as the System Design Technical Lead for IBM's revolutionary Neural Inference Accelerator for AI applications, NorthPole.
Dr. Filipp Akopyan was born in Moscow, Russia, where he grew up and completed school. Filipp started attending Rensselaer Polytechnic Institute (in Troy, NY) in September of 2001 and graduated in 3 years with a Bachelor's Degree in Electrical Engineering (with a minor in Economics) in May of 2004. Filipp was ranked #1 in the School of Engineering (Summa Cum Laude). His concentrations at RPI included Electronic Circuit Design and Signal Processing.
Filipp began a joint M.S. / Ph.D. program at Cornell University in September of 2004. At Cornell, he excelled at the Asynchronous VLSI (AVLSI) research group led by Professor Rajit Manohar. Filipp has authored several highly cited publications (over 8,000 total citations) in the areas of asynchronous design, tool flow architecture, signal processing, neuromorphic computing, 3D integrated circuits, and low-power chip design. Filipp is a recipient of numerous awards, patents and honors in the fields of Electrical and Computer Engineering.
After receiving his Ph.D. in Electrical and Computer Engineering (with a minor in Applied Mathematics) from Cornell University in 2011, Dr. Akopyan joined the IBM Brain-Inspired Computing Group led by Dharmendra Modha. During his IBM tenure, Filipp was one of the lead engineers, who created the world's most advance neuromorphic chip, TrueNorth, as part of the DARPA SyNAPSE program. He was also a key contributor on the DARPA Cortical Processor program. One of Filipp’s main contributions was leading the complex task of designing, simulating and verifying mixed Synchronous-Asynchronous circuits to implement the TrueNorth chip correctly and efficiently. His novel tool flow methodology has a potential of revolutionizing future state-of-the-art designs by making them extremely flexible and energy efficient. More recently, Filipp led a Team of Physical Design Engineers to successfully tapeout IBM's novel Inference Accelerator for AI applications, NorthPole, which is fully-functional on first silicone. At IBM Filipp has also been designing state-of-the-art energy-efficient Cognitive Systems and enabling Neural Algorithm development for various multi-sensory applications.
Google Scholar: http://scholar.google.com/citations?user=XkNtWmUAAAAJ&hl=en