Amazon today announced that the team from Emory University is the winner of the 2020 Alexa Prize.
“Congratulations to the team from Emory for their impressive work in making conversations between humans and Alexa more engaging,” said Prem Natarajan, Amazon vice president of Alexa AI's Natural Understanding organization. “We are delighted that, for the third year in a row, the winning team has set a new Alexa Prize record in terms of average ratings from users.”
In their own words
We interviewed some of the Alexa Prize Grand Challenge 3 participants (and one judge) about the competition, the role of COVID-19, and the future of socialbots.
The Alexa Prize, launched in 2016, is a competition for university students dedicated to advancing the field of conversational AI. Teams are challenged to design socialbots that Alexa customers can interact with via Alexa-enabled devices. Their ultimate goal is to meet the Grand Challenge: earn a composite score of 4.0 or higher (out of 5) from the judges, and have the judges find that at least two-thirds of their conversations with the socialbot in the final round of judging remain coherent and engaging for 20 minutes.
The Emory team, which created the Emora socialbot, is led by PhD student Sarah Fillwock, along with faculty advisor Jinho D. Choi. The team earned $500,000 for its first-place performance. Their work, along with that of the other Alexa Prize participants, is now outlined in a research paper posted to alexaprize.com.
While none of this year’s teams met the Grand Challenge, each team showed impressive progress toward that goal. Emora, the Emory University chatbot, earned first place with a 3.81 average rating. Stanford University’s Chirpy Cardinal team—led by PhD students Ashwin Paranjape and Abigail See, along with faculty advisor Christopher Manning—attained second place, and a $100,000 prize, with a 3.17 average rating. The third place team was Alquist, the socialbot from Czech Technical University. That team, led by PhD student Jan Pichl and faculty advisor Jan Sedivy, earned a $50,000 prize, with a 3.14 average rating for its socialbot.
In July 2019, 10 university teams were selected from a global pool of applicants to participate in the Alexa Prize Socialbot Grand Challenge 3. In May of this year, five of those teams came one step closer to reaching their goal.
The bots evolved throughout the competition to cover a wide range of topics, from sports and movies to current events—including in-depth handling of questions about the COVID-19 pandemic. They’ve also improved from previous competitions by getting funnier, finding ways of connecting with Alexa customers more effectively, and by retrieving more relevant information to conversation topics. Over the course of the competition, Alexa customers held more than 240,000 hours of conversations with socialbots, spanning tens of millions of interactions.
“The work by this year’s teams during a global pandemic was especially challenging and inspiring,” Natarajan said. “We are grateful for the work done by all of the teams competing in this challenge, for their creativity, passion, and contributions to both AI science, and to delighting customers.”
Check in with the Amazon Science website in the coming months to learn about application details and deadlines for the Alexa Prize Grand Challenge 4.