I braved the commute into and out of Boston yesterday to attend an incredible conference at MIT. One of the greatest things about being a journalist is being able to get a press pass to attend this type of event.
I was mostly interested in the morning session on artificial intelligence (AI), because I’m working on a book idea on this subject. The presenters were Eric Horvitz of Microsoft and Yann LeCun of Facebook.
Some highlights from Eric Horvitz’ talk:
- Artificial Intelligence today is about reasoning under uncertainty
- I was passionate to know what was going on with human brains. Biology had discovered how to do some beautiful magic that we had very little understanding of.
- Richer models will bring together multiple competencies and play a symphony of sorts – a symphony of intelligence.
- Deep competencies are what we perceive as conscious experience.
At the end of the talk, I was one of three audience members chosen to ask a question. My question:
“You mentioned that neural networks are very loosely inspired by biology. What are the major similarities and differences between what happens in a deep learning system and the brain?”
Eric Horvitz had a great answer, which I summed up in a tweet.
- You don’t need to know exactly how a feather works to build an airplane. We’re looking for the equivalent of aerodynamics for AI.
The rest of the day included sessions about urban planning and the future of energy. I especially enjoyed a talk on batteries and one on nuclear reactor designs. The weather was gorgeous — here I am with the view of Boston from the balcony.
