Skip to main content

Four Way Interview - Hector Levesque

Hector Levesque is Professor Emeritus in the Department of Computer Science at the University of Toronto. He worked in the area of knowledge representation and reasoning in artificial intelligence. He is the co-author of a graduate textbook and co-founder of a conference in this area. He received the Computers and Thought Award in 1985 near the start of his career, and the Research Excellence Award in 2013 near the end, both from IJCAI (the International Joint Conferences on Artificial Intelligence). His latest title is Common Sense, The Turing Test, and the Quest for Real AI.

Why computer science?

Computer science is not really the science of computers, but the science of computation, a certain kind of information processing, with only a marginal connection to electronics. (I prefer the term used in French and other languages, informatics, but it never really caught on in North America.) Information is somewhat like gravity: once you are made aware of it, you realize that it is everywhere. You certainly cannot have a Theory of Everything without a clear understanding of the role of information. 

Why this book?

AI is the part of computer science concerned with the use of information in the sort of intelligent behaviour exhibited by people. While there is an incredible amount of buzz (and money) surrounding AI technology these days, it is mostly concerned with what can be learned by training on massive amounts of data. My book makes the case that this is an overly narrow view of intelligence, that what people are able to do, and what early AI researchers first proposed to study, goes well beyond this.

What's next?

I have a technical monograph with Gerhard Lakemeyer published in 2000 by MIT Press on the logic of knowledge bases, that is, on the relationship between large-scale symbolic representations and abstract states of knowledge. We are working on a new edition that would incorporate some of what we have learned about knowledge and knowledge bases since then. 

What's exciting you at the moment?

For me, the most exciting work in AI these days, at least in the theoretical part of AI, concerns the general mathematical and computational integration of logical and probabilistic reasoning seen, for example, in the work of Vaishak Belle. It's pretty clear to all but diehards that both types of knowledge will be needed, but previous solutions have been somewhat ad hoc and required giving up something out of one or the other.

Comments

Popular posts from this blog

Models of the Mind - Grace Lindsay *****

This is a remarkable book. When Ernest Rutherford made his infamous remark about science being either physics or stamp collecting, it was, of course, an exaggeration. Yet it was based on a point - biology in particular was primarily about collecting information on what happened rather than explaining at a fundamental level why it happened. This book shows how biologists, in collaboration with physicists, mathematicians and computer scientists, have moved on the science of the brain to model some of its underlying mechanisms. Grace Lindsay is careful to emphasise the very real difference between physical and biological problems. Most systems studied by physics are a lot simpler than biological systems, making it easier to make effective mathematical and computational models. But despite this, huge progress has been made drawing on tools and techniques developed for physics and computing to get a better picture of the mechanisms of the brain. In the book we see this from two directions

The Ten Equations that Rule the World - David Sumpter ****

David Sumpter makes it clear in this book that a couple of handfuls of equations have a huge influence on our everyday lives. I needed an equation too to give this book a star rating - I’ve never had one where there was such a divergence of feeling about it. I wanted to give it five stars for the exposition of the power and importance of these equations and just two stars for an aspect of the way that Sumpter did it. The fact that the outcome of applying my star balancing equation was four stars emphasises how good the content is. What we have here is ten key equations from applied mathematics. (Strictly, nine, as the tenth isn’t really an equation, it’s the programmer’s favourite ‘If… then…’ - though as a programmer I was always more an ‘If… then… else…’ fan.) Those equations range from the magnificent one behind Bayesian statistics and the predictive power of logistic regression to the method of determining confidence intervals and the kind of influencer matrix so beloved of social m

How to Read Numbers - Tom Chivers and David Chivers *****

This is one of my favourite kinds of book - it takes on the way statistics are presented to us, points out flaws and pitfalls, and gives clear guidance on how to do it better. The Chivers brothers' book isn't particularly new in doing this - for example, Michael Blastland and Andrew Dilnot did something similar in the excellent 2007 title The Tiger that Isn't - but it's good to have an up-to-date take on the subject, and How to Read Numbers gives us both some excellent new examples and highlights errors that are more common now. The relatively slim title (and that's a good thing) takes the reader through a whole host of things that can go wrong. So, for example, they explore the dangers of anecdotal evidence, tell of study samples that are too small or badly selected, explore the easily misunderstood meaning of 'statistical significance', consider confounders, effect size, absolute versus relative risk, rankings, cherry picking and more. This is all done i