Monday, January 6, 2014

Simple Science: Social Robotics

This post is the first in our Simple Science Descriptions series of pieces written with the Up-Goer Five Text Editor restricting writers to the ten hundred most used words in English. It's harder than you might think! Send your entries (preferably under 350 words) to jeanyang [at] mit [dot] edu.


Humans are very good at some things, like recognizing things and using words. Computers are very good at other things, like working very fast with large numbers. I try to make the computers good at the things that humans are good at, like talking to and understanding people.

I build computers with bodies that can see, hear, and move around. Some of these things are already in the world - they build cars or clean your floor. But they aren't very good at working with people, because they don't understand how people act.

I try to make these seeing, hearing, moving computers better at dealing with humans by helping them understand people's voices and faces and also by making the computers act and move in ways that make people feel safe. This has many practical uses, because these computers can then do things that humans had to do before, even if they didn't want to (like making the same things over and over again all day).

Seeing, hearing, and understanding other people is very hard, even though it seems easy for us. So my work can also help us understand how humans do these things so easily by letting us check whether our ideas about how humans hear, see, and understand others are right. We make a guess about how it works in humans, then check to see if our guess is right by making the computer do what we think humans do and seeing if the computers act the way humans do.

- Sam Spaulding, graduate student

No comments:

Post a Comment