An interview with Google's Head of Research

Posted at: Wednesday 12th September 2007 by Sarah Lewthwaite

We recently got the chance to chat to Dr Peter Norvig, Google's Head of Research about where the big G is headed - there's no interest in hardware apparently, but computer vision, face recognition and making sense of the narrative of search are all on the agenda

Laid back and in a loud shirt, Dr Peter Norvig looks like the kind of academic normally seen in Hollywood movies. It's not surprising that he's assured and relaxed, though - not only is he a well regarded researcher at the top of his game - he's also the Director of Research at one of the web's most important and wealthiest companies.

Google.

Dr Norvig was in the UK for the Association For Learning Technology conference (ALT-C). At a pre-conference workshop, a collection of delegates brainstormed potential areas for expansion on the basis of current Google Apps. One academic asked what payment the think tank can expect, to which Norvig quickly responded: 'Free internet searches for life.'

In this sense, Norvig is amiable, but politic. He has been at Google since 2001 and Director of Research since 2006. With over 50 Computer Science publications in artificial intelligence, natural language processing and software engineering to his name, Norvig has literally authored text books. We managed to get some time with him at the conference to quiz him about Google and his views on current and future technologies.

Custom PC: What are the big projects at Google research at the moment?

Dr Peter Norvig: We’re looking at a lot of different things. The biggest projects we have right now are in machine translation and speech recognition, computer vision and face recognition, team recognition and so on. We’re looking at the areas that are peripheral to the main things we're doing: trying to understand text and pushing that out…But we do have many projects that are closer to more traditional work for us; language understanding, work on advertising, network and theoretical economics, questions like ‘what’s the right way to run an auction?’ and also more practical things, say,‘Gee, if I could squeeze a 1% improvement to the way we attach ads to queries, well, that’s worth $100 million dollars’.

I blink and move on to my next question.

CPC: Obviously Google is a commercial organisation but it’s also famed for its philanthropic edge with its 'Don’t be Evil' motto. Is it difficult balancing those demands when it comes to research?

Dr Norvig: No. One of the things that’s great about Google is that the focus is on the consumer. Mainly we’re here to make the consumer happy. One of the things that consumers do when they’re happy is view ads, and we get paid for that. We take this seriously and make a good job of it.

We don’t really have to do strict accounting in each division. In many companies you see a lot of divisiveness: 'My business unit has to achieve so much per year and this other competing unit is trying to one-up me.' You see different divisions of the company competing with each other rather than helping each other out. At Google we’ve never had that because we know where our money is coming from and we know we’ll get our share. Individual units at Google don’t have to worry about making money, they don’t all have to be profitable. Most of them are just there to make the consumer happy and if they’re happy then the rest of the company is happy.