Professor Stefano Fusi at Center for Theoretical Neuroscience at Columbia University (who was part of IBM Team for DARPA SyNAPSE in Phases 0, 1, and 2) has released a very interesting pre-print entitled “Energy-efficient neuromorphic classifiers”. Here is the Abstract (highlights are mine):
Neuromorphic engineering combines the architectural and computational principles of systems neuroscience with semiconductor electronics, with the aim of building efficient and compact devices that mimic the synaptic and neural machinery of the brain. Neuromorphic engineering promises extremely low energy consumptions, comparable to those of the nervous system. However, until now the neuromorphic approach has been restricted to relatively simple circuits and specialized functions, rendering elusive a direct comparison of their energy consumption to that used by conventional von Neumann digital machines solving real-world tasks. Here we show that a recent technology developed by IBM can be leveraged to realize neuromorphic circuits that operate as classifiers of complex real-world stimuli. These circuits emulate enough neurons to compete with state-of-the-art classifiers. We also show that the energy consumption of the IBM chip is typically 2 or more orders of magnitude lower than that of conventional digital machines when implementing classifiers with comparable performance. Moreover, the spike-based dynamics display a trade-off between integration time and accuracy, which naturally translates into algorithms that can be flexibly deployed for either fast and approximate classifications, or more accurate classifications at the mere expense of longer running times and higher energy costs. This work finally proves that the neuromorphic approach can be efficiently used in real-world applications and it has significant advantages over conventional digital devices when energy consumption is considered.
World Economic Forum named “Neuromorphic technology” as one of “Top 10 Emerging Technologies of 2015” and specifically cited IBM’s TrueNorth Chip (see page 12 of the report).
Guest Post by Ben G. Shaw, Organizing Chair of Cognitive Systems Colloquium.
This is continued from the previous post dated November 12, 2014.
To highlight the transformative potential of IBM’s Neurosynaptic System and its impact on computation in the Cognitive Era, IBM Research hosted nearly 200 eminent thinkers and pioneers in the field of brain-inspired computing at the IBM Research – Almaden Cognitive Systems Colloquium. The program featured over a dozen outstanding speakers and distinguished panelists. Attendees included nearly 200 thought leaders and potential early adopters from government, industry, academia, research and the venture community.
Recurring Themes of the Day:
- The Brain: how advances in understanding nature’s most efficient and powerful computational substrate are revealing new paradigms for computing
- Technology: as von Neumann computation comes up against fundamental limitations that are bringing Moore’s law to an end, how new approaches can revolutionize important classes of computation
- Applications: how efficient, embedded neural computation may benefit individuals, businesses and society by making objects, environments and systems more aware and responsive
- Ecosystems: how new technologies and offerings will gain breadth, depth and momentum to transform industries from robotics to healthcare, agriculture to mobile devices, transportation to public safety.
SyNAPSE Deep Dive:
In addition to reviewing the state of knowledge in the field of brain-inspired computing and a forward-looking panel discussion, participants took a concentrated “Deep Dive” into the recently announced IBM Neurosynaptic System including the 1-million neuron TrueNorth chip, architecture, development boards, programming paradigm, applications, education and ecosystem. Inspired by the brain, TrueNorth is an architecture and a substrate for non-von Neumann, event-driven, multi-modal, real-time spatio-temporal pattern recognition, sensory processing and integrated sensor-actuator systems. TrueNorth’s extreme power efficiency and inherent scalability will revolutionize applications in mobile and embedded systems, at the same time allowing neural algorithms to achieve previously unattainable scales, running quickly, efficiently and natively in hardware.
- Brain-inspired Computing: A Decade-Long Journey (20:27)
- Part I: The Need for a New Architecture, TrueNorth & Compass, Transduction, Live demos (25:45)
- Part II: Architecture, Neuron, Training for TrueNorth, MNIST Example (24:10)
- Part III: Corelet Development, Corelet Programming, Hardware Placement (18:27)
- Part IV: Mobile Deployment, Scale Deployment, SyNAPSE University (14:49)
Distinguished Speakers and Panelists:
- From BrainScales to the Human Brain Project: Neuromorphic Computing Coming of Age (24:33)
Karlheinz Meier, Professor & Co-Director, Human Brain Project, University of Heidelberg
- Brain-inspired Computing: A Decade-Long Journey (20:27)
Dharmendra Modha, IBM Fellow and Principal Investigator, IBM SyNAPSE Program
- Cell Type and Computation (17:11)
Michael Hawrylycz, Investigator, Allen Institute for Brain Science
- Synesthesia’s Challenge to Brain-Inspired Computing (17:24)
Richard Cytowic, Author of Synesthesia: A Union of the Senses
- Visual Cortex in Silicon (26:30)
Vijaykrishnan Narayanan, Professor, Penn State University and PI, NSF Expeditions in Computing
- Silicon Retinas (17:18)
Tobi Delbruck, Co-Founder, INILabs, and Professor, ETH Zurich
- Asynchronous Circuits (21:54)
Rajit Manohar, Professor, Cornell Tech
- A Quest for Visual Intelligence (23:10)
Fei-Fei Li, Professor and Director, AI Lab, Stanford University
- Panel: Brain, Computers, Society, Future
- Andreas Andreou, Professor, Johns Hopkins University
- Gary Marcus, Professor, NYU
- Horst Simon, Deputy Laboratory Director, Lawrence Berkeley National Laboratory
- Jayashree Subrahmonia, Vice President for Products, IBM Watson Group
- Jim Spohrer, IBM Director of Global University Programs (Moderator)
- Mark Anderson, CEO, Strategic News Service and Chair, Future in Review Conference
- Miyoung Chun, Executive Vice President for Science Programs, The Kavli Foundation
The audience included luminaries such as Turing Prize Awardee, Ivan Sutherland, and Von Neumann Prize Awardee, Nimrod Megiddo. Four IBM Fellows were in attendance (Ronald Fagin, C. Mohan, Hamid Pirahesh, Stuart Parkin), as were prominent founders and visionaries in the field of brain-inspired computing, including Warren Hunt (UT Austin), Tim Lance (NYSERNet), Einar Gall (Neurosciences Institute), Gert Cauwenberghs (UCSD), Ken Kreutz-Delgado (UCSD) and Jeff Krichmar (UC Irvine).
On March 10, 2015, at the 56th Foundation Day of IIT Bombay, I was selected for Distinguished Alumnus Award. I am grateful for the education that I received at IIT Bombay, for my teachers, for my fellow students, for my hostel mates, for the mess workers who fed me for four years, for the support staff, for my colleagues at IBM, and, of course, my family. Of the nearly 50,000 alumni, to date, roughly 100 have been honored. Previous Awardees include Nandan Nilekani and Kanwal Rekhi as well as two IBM Fellows Subramanian Iyer and Ramesh Agarwal.
Photo Credit: Hita Bambhania-Modha