“How Computers Learn”
Peter Norvig demonstrates how machines can independently learn to perform complex tasks or develop sophisticated game strategies.
Peter Norvig is a Director of Research at Google Inc. Previously he directed Google’s core search algorithms group and was NASA’s senior computer scientist. He is co-author of Artificial Intelligence: A Modern Approach, the leading textbook in the field, and co-teacher of an Artificial Intelligence class that signed up 160,000 students. He is a fellow of the AAAI, ACM, California Academy of Science, and American Academy of Arts & Sciences.
On Thursday, 26 March 2015, the third edition of our Vienna Gödel Lecture series took place - Peter Norvig presented how computers learn. The lecture met with a massive response from the audience, which flocked in droves to the three lecture halls. Peter Norvig is a leading expert in the field of Artificial Intelligence and author of the standard work Artificial Intelligence: A Modern Approach.
Models Are Wrong, But Some Are Useful
Peter Norvig likes this quotation from British statistician George Box a lot - and explained based on different models how computers can independently learn to perform complicated tasks using training data: translate texts, recognize the content of pictures or videos or develop sophisticated game strategies. Computer scientists have developed complex programming languages, but there are limitations to these techniques, too. “This idea of machine learning is now everywhere, ” says Norvig. “And we are doing it all by training computers rather than programming them.“ He vividly described how this learning process could deliver amazingly accurate results using simple principles and supported by a large amount of training data. After the lecture, Norvig answered numerous questions of the interested audience in detail.
You have missed the Vienna Gödel Lecture with Peter Norvig? Then watch it here.