Get to Know Your Fellow Researchers: Geoffrey Zweig

What sparked your interest in speech and language processing?

I have always been amazed that people speak very differently, and yet we can understand each other. I find it tremendously fascinating that there is a pattern that is constant across all the different ways that people say the same words and sentences. Not only that, we can do it even when there’s music and noise, and a lot going on in the background. I think I was first amazed by this when I was 14 or 15.

How do you think speech and language processing is changing the society for the new generation?

I think artificial intelligence will slowly seep into more and more aspects of our lives. Right now we can ask our cell phones for almost any kind of information and get it. As time goes on, agents like Cortana and Siri will become increasingly able to understand the details of our lives and interact with us in more and more useful and entertaining ways.

What is your holy grail in speech and language processing? When will we achieve it?

The cocktail party problem. Imagine sprinkling some microphones around a room, and being able to recognize people talking at a party, with lots of people talking, music, glasses and plates clinking, and so on. We still are not there yet. This is different than language processing, though. For that, perhaps we could get an automated agent to hold up it’s end of a conversation in the party.

Do you have any specific advice for students, junior faculty or others early in their careers?