Not many people can say that they have communicated with robots, worked for Disney, and graced the stages of both Jeopardy! and Who Wants to be a Millionaire?. But for Samuel Spaulding, a senior computer science major in Jonathan Edwards, this is simply the beginning of a promising future in social robotics.
As a member of his high school math and Quiz Bowl teams, Spaulding’s interest in science started early. His interest in artificial intelligence, specifically, was sparked during his senior year after reading the book Gödel, Escher, Bach: An Eternal Golden Braid. However, it was only when he came to Yale that he became interested in computer science.
“When I first got here, I’d never taken a computer science course,” Spaulding recalled. “But freshman year I took this Intro Programming class, really loved it, and thought ‘Okay, computer science is something that I can really get into.’” Drawing inspiration from his interest in cognitive science, Spaulding eventually found a field that combined both of his interests — applying artificial intelligence to robotics.
At the Yale Social Robotics lab, Spaulding’s research involves finding ways for humans and robots to interact comfortably with one another through spoken language. “The idea first came when I was thinking, ‘How can you make interactions between humans and robots more natural?’” Spaulding said. “Language is the most natural way that humans communicate. It’s what we use to communicate with each other. Speech technology like voice recognition and synthesized speech weren’t good enough in the past to justify working with, but they have been improving so rapidly that it’s feasible to think now about this type of interaction.” Thus, one of Spaulding’s goals is to teach robots about the environment around them using speech.
In a project that culminated in his senior thesis, Spaulding created a robotic system that adjusts its behavior depending on the verbal input it gets from the user. After hearing the user describe an object, the robot can identify the object and establish a sentiment score based on what the user says about it. Something that the user describes in a positive way would therefore be assigned a higher sentiment score than something described negatively. Afterwards, the user will ask the robot about the object at hand. For example, the robot may be asked, “How do you feel about carrots?” The robot then selects behaviors to execute that are socially appropriate, given the sentiment score it has learned for carrots from the user’s description. By teaching robots how to react to objects that they encounter, this undertaking provides a step towards improving human-robot communication.
In addition to working in the Social Robotics Lab, Spaulding has been putting his computer science skills to use in various ways. In the summer of 2011, he worked at Amazon to develop a website that assessed team performance. The experience gave him valuable programming experience and showed him how a standard software engineering company functions on a grand scale. Furthermore, he worked on a team to create an Android app called SmileIKnow, which became a finalist at the Amazon Mobile Security Hackathon. The following summer, Spaulding worked as a research assistant for Walt Disney Imagineering. “I was doing artificial intelligence research for tools that might eventually see use in the parks,” he said.
When Spaulding is not working on robots, he may be found reading, playing video games, or enjoying the outdoors. As a trivia enthusiast, he is also on Yale’s Quiz Bowl Team. “I like to keep my skills sharp with bar trivia, go out and play on a team there,” said Spaulding. “It’s a lot of fun.” And it pays off: he placed second in the Jeopardy! College Championship.
Given his wide range of experiences, Spaulding had a lot to talk about when selected to speak at the TedxYale conference this year. Drawing upon his extensive background with artificial intelligence, he steered away from the purely technical aspects of his research and instead explored the unique nature of human intelligence.
Next year, Spaulding plans to continue his education in graduate school, although he has not yet committed to any particular institution. Indeed, with all that he has achieved in the past four years, Spaulding will certainly have a busy and rewarding post-Yale career ahead of him. “I’ll be going to graduate school for robotics, so hopefully I’ll have a lot of opportunities there,” he said. “I’ve got some ideas for things that I’d like to work on.”