2015 Rockwood Memorial Lecture

INC sponsors the H. Paul Rockwood Memorial Lectureship held annually. The Rockwood Memorial Lectureship Fund was gifted to the Institute by Mr. and Mrs. Jerome Rockwood in memory of their late son's interest, studies, and work in the neural computation field.
The Rockwood Memorial Lectures are endowed by Mr. and Mrs. Jerome Rockwood in memory of their late son, Paul, who received a B.S. in Computer Science from UCSD in 1980 and then obtained a second degree B.A. in Psychology in 1981. In 1983 he started a company, Integral Solutions, to develop a universal language translation, but died tragically in a mountaineering accident before he could fulfill his promise.


Neuromorphic Chips: Combining Analog Computation with Digital Communication

Dr. Kwabena Boahen
Associate Professor, Bioengineering
Brains in Silicon Lab
Stanford University

4:00 - 5:30, Friday, April 24, 2015

San Diego Supercomputer Center Auditorium, Floor B-2, 10100 Hopkins Dr., La Jolla, CA 92093



I'll argue that combining analog computation with digital communication can effectively combat increased heterogeneity (a.k.a., mismatch) and stochasticity (a.k.a., noise) as transistors scale down to a few nanometers. Intriguingly, the brain adopts precisely such a hybrid approach to combat the heterogeneity and stochasticity that arise from its nanoscale ion-­-channels' probabilistic expression and operation. Neurons' dendritic trees combine thousands of graded potentials to produce an output (c.f., analog computation) that their axonal arbors transmit to thousands of other neurons in the form of all-­-or-­-none action potentials (c.f., digital communication).

To support my argument, I'll present a Kalman-­-filter-­-based brain-­-machine interface and a three-­-degree-­-of-­-freedom robot-­-arm controller implemented on a neuromorphic chip that combines analog computation with digital communication much like the brain does. The neuromorphic chip's spiking-­-neuron network was configured to perform these operations using the Neural Engineering Framework (NEF), a formal method for approximating arbitrary nonlinear dynamical systems with networks of spiking neurons. Using NEF, you can, for example, get a spiking-­-neuron network to perform temporal integration or rotate an n-­-dimensional vector by a given angle. NEF proved robust to heterogeneity introduced by transistor mismatch as well as stochasticity introduced by probabilistic spike delivery. As an added bonus, NEF produced interconnection-­-weight matrix with low-­-rank—it could be stored using an order-­-of-­-magnitude less memory.

Bio: Kwabena Boahen received the B.S. and M.S.E. degrees in electrical and computer engineering from the Johns Hopkins University, Baltimore, MD, both in 1989 and the Ph.D. degree in computation and neural systems from the California Institute of Technology, Pasadena, CA, in 1997. He was on the bioengineering faculty of the University of Pennsylvania from 1997 to 2005, where he held the first Skirkanich Term Junior Chair. He is presently a Professor in the Bioengineering Department of Stanford University, with a courtesy appointment in Electrical Engineering. He directs Stanford's Brains in Silicon Laboratory, which develops silicon integrated circuits that emulate the way neurons compute, linking the seemingly disparate fields of electronics and computer science with neurobiology and medicine.

Prof. Boahen's contributions to the field of neuromorphic engineering include a silicon retina that could be used to give the blind sight, a self-­-organizing chip that emulates the way the developing brain wires itself up, and a specialized hardware platform (Neurogrid) that simulates a million cortical neurons in real-­-time—rivaling a supercomputer while consuming only a few watts. He has received several distinguished honors, including a Fellowship from the Packard Foundation (1999), a CAREER award from the National Science Foundation (2001), a Young Investigator Award from the Office of Naval Research (2002), a Pioneer Award from the National Institutes of Health (2006), and a Transformative Research Award from the National Institutes of Health (2011). His 2007 TED talk, "A computer that works like the brain", has been viewed half-­-a-­-million times.


Organized by:
UCSD Institute for Neural Computation


Past H. Paul Rockwood Memorial Lectureships

  • 2014 Lecturer: Dr. Bruce McNaughton,
    "Doughnuts in the Brain: A Toroidal Attractor Theory of the Cognitive Map"

  • 2013 Lecturer: Dr. Dmitri B. Chklovskii,
    "Can connectomics help us understand neural computation? Insights from the fly visual system "

  • 2012 Lecturer: Dr. Eugene M. Izhikevich,

  • 2011 Lecturer: Rodney Douglas,
    "Constructive Cortical Computation"

  • 2010 Lecturer: Geoffrey Hinton,
    "Deep learning with multiplicative interactions”

  • 2009 Lecturer: Josh Bongard,
    "Investigations at the Interface of Morphology, Evolution and Cognition”
    "Resilient Machines (Through Continuous Self-Modeling)"

  • 2008 Lecturer: Jeff Hawkins, Founder, Palm, Inc. Handspring, and Numenta,
    "Computing Beyond Turing: How neocortical theory is shaping the future of computing"

  • 2007 Lecturer: Tomaso Poggio, Center for Biological and Computational Learning, Computer Science and Artificial Intelligence Laboratory, McGovern Institute for Brain Research, Massachusetts Institute of Technology,
    "What Should Computer Vision Learn from Neuroscience?"

  • 2006 Lecturer:Harvey Karten, Department of Neurosciences, UCSD
    "Unwiring the Brain: Neuromorphic Engineering of Motion Detection"
    View Movie (Windows Media Player / 100MB): Part1 - Part2 - Part3

  • 2005 Lecturer: Jeffrey L. Elman, UC San Diego,
    "Generalizing beyond our experience: Lessons from neural networks"

  • 2004 Lecturer: Richard Gregory, University of Bristol,
    "A Periodic Table for Perception"

  • 2003 Lecturer: Stephen Wolfram, Wolfram Research, Inc.,
    "A New Kind Of Science"

  • 2002 Lecturer: Pasko Rakic, Yale University,
    "Building the Cerebral Cortex: From Stem Cells to Complex Architecture"

  • 2001 Lecturer: Michael Dickinson, UC Berkeley,
    "How Fruit Flies flap for Flight Forces: Neuro-Mechanics of a Complex Behavior"

  • 2000 Lecturer: Christoph von der Malsburg, Bochum University and University of Southern California,
    "The Learning Problem"

  • 1999 Lecturer: Dana Ballard, University of Rochester,
    "Single-Spike Models of Predictive Coding"

  • 1998 Lecturer: Geoffrey Hinton, University of Toronto,
    "Finding Structure in Ensembles of Images"

  • 1997 Lecturer: Michael Jordan, Massachusetts Institute of Technology,
    "Graphic Models, Neural Networks, and Variational Methods"

  • 1996 Lecturer: Jack Cowan, University of Chicago,
    "Geometric Visual Hallucinations, Migraine Auras, and Visual Illusions: What They Tell Us About Visual Cortex Circuitry"

  • 1995 Lecturer: Geoffrey Hinton, University of Toronto,
    "Neural Networks that Learn by Generating Fantasies"

  • 1994 Lecturer: Stephen Grossberg, Boston University,
    "Neural Networks for Learning, Recognition, and Recall"

  • 1993 Lecturer: James L. McClelland, Carnegie-Mellon University,
    "The Interaction of Nature and Nurture in Development: A Parallel Distributed Processing Perspective"