http://www.neurdon.com/wp-content/uploads/2009/02/image2-271x300.jpg
271
300
Reverse engineering the brain?
In a recent invited talk at the Department of Cognitive and Neural Systems, Lloyd Watts, neuroscientist turned entrepreneur (founder, chairman and CTO of Audience Inc., a Silicon Valley company that commercializes technology derived from auditory neuroscience research), presented his “strategy” on how to go about a gargantuan task: reverse-engineering the brain. With a military strategy analogy, the problem is the following: what is the best way to occupy an enemy territory? Should the invading army occupy simultaneously the target territory from all its borders, or should all troops focus on a narrow strip of land, occupy it, consolidate the territory and exploit its resources, and then move on to the next target? Lloyd Watts, the neuroscientist-entrepreneur, seems to suggest that the second strategy is the winning one.
Riesenhuber and Poggio supplied a seminal model of object recognition in 1999. It derived a lot of its power from sheer simplicity. With just a few mathematical operations it seemed to model the entirety of the ventral stream, the area of the brain dedicated to processing "What" information, i.e. information about the identity of an object. It starts with a layer of Gaussian-tuned `simple' or S cells, which respond to particular line orientations. That is, a particular S cell might respond to a diagonal line in a particular spot in an image. Then, all S cells of the same orientation feed to a 'complex' or C cell, which represents the maximally activated S cell. In CS terms, they take an argmax over a local neighborhood.
The challenge of building, within a few decades, a computer chip on the scale of a patch of biological cortex is a race involving many labs in academics and industry around the world.
Humans are remarkably good at identifying the same face across illuminations, positions, deformations, and depths. The same face can even be identified through fences, glass, and water. The possible number of contexts for a face to appear in are infinite, yet we can identify it instantaneously. For whatever reason, we are really good at identifying objects, but researchers have struggled to make computers even semi-competent at it. One of the more valiant efforts is Yann LeCun's use of