One of the other background projects I've been working on is with Ken Zolot, who I teaches the i-Teams course, where student teams build commercialization plans for Deshpande Center innovations. One of the technologies that came through earlier is a method of visual recognition that attempts to simulate how neurons operate. Theoretically, this will provide faster photo/video recognition for internet applications than current methods that utilized pixel differentiation. The technology comes out of the Department of Brain and Cognitive Sciences, and the Lab of Tomaso Poggio:
Right now, a number of VC's are crowding around this technology to do a seed round. Most likely, it's greed caused by YouTube and the potential to ride off of that valuation. However, this is a crowded space, with a lot of competing methodologies. At this point, I'm encouraging the Post-Doc who's working on this to build some sort of conceptual demo that uses the core technology; it's important to understand how well this stuff really works. More on this as it progresses.
Comments