Monday, January 23, 2012

Big Wide World - It's been fun, just don't follow in my footsteps

Sean O'Neill, contributor
(Image: Microsoft Research)

Chris Bishop heads up a team at Microsoft Research Cambridge in the UK, where he holds the title Distinguished Scientist - the company's highest research position. He leads the Machine Learning and Perception group. 

What attracted you to machine learning?
The film 2001: A Space Odyssey, with the intelligent computer, HAL, piqued my interest when I was a child. Later, I thought it would be amazing to build an intelligent machine but the research going on at the time didn't seem to be going anywhere. Then, in the late 1980s, along came neural networks and statistical approaches, and this allowed us to teach a computer to become intelligent rather than program it to be so. Since that approach is much closer to how the brain works, I felt it was more likely to succeed, so I got involved. 
What do you see on the horizon for the field?
The big thing over the next 10 years will be helping computers to be more like humans, or come closer to humans. Take Microsoft's Kinect system for the Xbox 360, which tracks body movement. We see that not just as a way to control computer games, but also as the first step to natural user interfaces - that is, making it as easy to interact with a computer as it is to interact with a person. The computer will ultimately understand your body language, your gestures, your voice, your facial emotions and so on. 

What is it like working for a global company like Microsoft?
It is the best research environment I know. We have a great deal of freedom - it's not a perk, like free coffee, it's part of the job. We have so much freedom because we are paid to figure out the questions as well as the answers. Working at Microsoft also means that we can make an impact on the world much more easily than those in an academic research environment. Sure, in academia you can take out patents and form start-up companies, but we can have a clever idea and within 12 months it can be going out the door to 100 million people. I can make an impact at Microsoft - a clever idea can go out to 100 million people within a year. 

Your career path has been anything but straight...
It has meandered somewhat. As an undergraduate I read physics and my PhD was very abstract; I worked with Peter Higgs and David Wallace on quantum field theory. After that I wanted to do something very practical so I switched to work on magnetic confinement fusion, as part of the ongoing effort to develop fusion reactors. After about eight years, I got interes

Is it good to move around so much?
It's a trade-off. If you work in one specific area all of your life, you become a world expert. On the other hand, if you move fields you get to try new things, which I loved. Looking back, my advice to others would be don't chop and change too much. I wouldn't necessarily encourage anyone to follow in my footsteps, but it has been a lot of fun.

No comments:

Post a Comment