November 27, 2018
By Bonnie Dickson
UC Davis Team Gunrock won the global 2018 Amazon Alexa Prize for creating the best conversational chatbot and advancing modern artificial intelligence. The announcement of this $500,000 prize was made yesterday at the annual Amazon Web Services re:Invent conference in Las Vegas.
Zhou Yu, an assistant professor in the College of Engineering’s Department of Computer Science, led the 10-student team to victory. She joined UC Davis in 2017 and was recognized in FORBES’ 2018 “30 Under 30” in Science list for her work developing algorithms that enable software to adapt to users.
“When our team first came together, we were competing against teams that had already participated in this challenge,” Yu said. “For us, it was our first time competing. Now, we are the best in the world in social conversational systems. They really came together and made this happen.”
The team’s bot – affectionately named “Gunrock” after the university’s mascot – earned a final score of 3.1 out of 5. Gunrock was also able to maintain an average of 9 minutes and 59 seconds of conversation during the competition’s final round of judging. Team Alquist from the Czech Technical University in Prague earned second place with a score of 2.6.
The competition was the culmination of nearly a year of research and development efforts by Team Gunrock. The team programmed their conversational bot using data from millions of Amazon Alexa user conversations.
Some of the distinguishing features of Gunrock included incorporating language disfluencies, or humanlike pauses such as “hm,” or “ah.” These humanlike attributes help distinguish Gunrock from traditional artificial intelligence bots.
The team also designed a robust natural language understanding model for their bot, which helped Gunrock break down dialogue into self-contained semantic units and parse through language to better determine context. They also integrated structured knowledge bases like Google Knowledge into the bot. This helped Gunrock handle a wide variety of user behaviors, such as topic switching and question answering.
First-year computer science graduate student Chun-Yen (Arbit) Chen led the UC Davis team, with Yu as the team’s faculty adviser. Other students on the team were: Ashwin Bhandare, Antara Bhowmick, Austin Chau, Shreenath Iyer, Kevin Jesse, Giritheja Sreenivasulu, Weiming Wen, Yi Mang Yang, Dian Yu, and Mingyang Zhou. The team included 10 graduate students and one undergraduate student.
Although they have not yet decided how to use their prize money, the team will definitely compete again next year, Yu said.
The competition’s formal proceedings are published online.