Actions:
Metadata:
------
Account:
Login
  • Deme account
  • OpenID


Navigation
Recently Viewed
SSP Forum: Pentti Kanerva » Show

Nov. 28, 2011

Start Date

& End Date

12:15 p.m.

Start Time

to

1:05 p.m.

End Time

at 460-126

Location

Symbolic Systems Forum
presents
"Hyperdimensional Computing for Modeling How Brains Compute"
Pentti Kanerva
Redwood Center for Theoretical Neuroscience, UC Berkeley, and CSLI

Monday, November 28, 2011
12:15-1:10 pm
Building 460, Room 126
map link: http://campus-map.stanford.edu/index.cfm?ID=01-460

ABSTRACT:

We think of the brain as a kind of a computer, but what
brains accomplish is very different from what we can do with
computers.  It is very likely that our present computer
architecture is incapable, in principle, of producing
brainlike behavior.  We can look at neuroanatomy for ideas
about computer architecture with more brainlike powers.
Striking about the brain's circuits is their very size,
which in computer terms means that we should compute with,
say, 10,000-bit words rather than with 16-to-64-bit words.
What would computing with such wide words be like?
Neural-net associative memories (e.g., Willshaw, Hopfield)
are early examples of that kind of computing.  They are
content-addressable and work with incomplete or noisy data
vectors.  High-D vectors' tolerance for noise is well know.
Less well known is their suitability for symbolic computing.
The required operations form the core of a new kind of
computing that seems particularly suited for modeling
cognition, as well as for implementation in nanotechnology.
The best-known model of the kind, and perhaps the first,
is Plate's Holographic Reduced Representation in the early
1990s, but the field remains largely unexplored.  Language
provides fertile grounds for the exploration.

REFERENCE

Kanerva, P.  Hyperdimensional Computing: An Introduction
  to Computing in Distributed Representation with
  High-Dimensional Random Vectors.  COGNITIVE COMPUTATION
  1(2):139-159, 2009.

  http://www.springerlink.com/content/966151841g415165/
  http://redwood.berkeley.edu/wiki/Publications  (free)

Description

Desktop Browser | RSS | Increase text size | Decrease text size | Fullscreen | Dictionary/Thesaurus | Translate

Copyright ©2008-2013 to Deme | Source Code | Authors | Admin | Recent Changes

For Internet Explorer users: Click on the Tools menu, located at the top of your browser window. When the drop-down menu appears, select the option labeled Full Screen.

For Chrome users:Click on the Chrome "wrench" icon, located in the upper right hand corner of your browser window. When the drop-down menu appears, select the choice labeled Full Screen.

For Firefox user:Click on the View menu, located at the top of your browser window. When the drop-down menu appears, select the option labeled Full Screen.

For Safari users: Safari currently does not support the ability to go fullscreen.