MIT plans AI redux

THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY (MIT) is looking to re-ignite the development of artificial intelligence (AI) by 'going back and fixing mistakes', something that sounds like it came straight out of a movie. Sadly, however, no one is planning to send Tom Cruise back in time.

Today in a report, the MIT news office said that a $5 million project called the Mind Machine Project, or MMP, has been launched with the aim of starting over on research and development that first started 50 years ago. A loose collective of some twenty four experts including professors, researchers, students and postdocs will work together for five years to create new intelligent machines that will live up to the dream of AI.

Having said that, it all sounds a little wooly. Neil Gershenfeld, one of the leaders of MMP, added, "whatever that means. Essentially, we want to rewind to 30 years ago and revisit some ideas that had gotten frozen." He also said that the five year timeline might be constrictive, but added, "We need good challenging projects that force us to bring our program together."

There is much to be done, other members explained. "Considering the outrageous optimism of much of the early hype for AI, it is no wonder that it couldn't deliver. This is an occupational hazard of many new fields," said Daniel Dennett, a professor of philosophy at Tufts University. "The reality is not dazzling, but still impressive, and many applications of AI that were deemed next-to-impossible in the '80s are routine today."

In order to 'fix' AI the team will look at three areas of its development - the mind, memory and the body - and consider how to apply them to artificial intelligence. According to Gershenfeld, research has become stuck in these areas and needs re-igniting. "How do you model thought?" he asked. "What's been missing is an ecology of models, a system that can solve problems in many ways."

We are glad that no one has asked us to join in as the whole thing sounds confusing to us.

"The pieces are very disparate; they're not necessarily built in a compatible way. There's a similar pattern in AI research. There are lots of pieces that work well to solve some particular problem, and people have tried to fit everything into one of these," continued Gershenfeld. "Instead of searching for silver bullets, we're looking at a range of models, trying to integrate them and aggregate them".

The full report is here.

In the meantime, we all want to marry models. Keep on living the dream, fellas. µ