Aucun résultat trouvé
February 01, 2018Ashwin Ram
I’m pleased today to announce the 2018 teams selected to compete for the Alexa Prize, a $3.5 million university challenge to advance conversational AI. The teams will create socialbots that can converse coherently and engagingly with humans on a range of current events and popular topics such as entertainment, sports, politics, technology, and fashion.
Since early December, we’ve received applications from leading universities from more than 15 countries. Each application was carefully reviewed by senior Amazon personnel against rigorous criteria including scientific contribution, technical merit, novelty, and ability to execute.
The eight teams selected to compete, listed below in alphabetical order, are:
Each team will receive a $250,000 research grant, Alexa-enabled devices, free Amazon Web Services (AWS) to support their development efforts, access to new Alexa Skills Kit (ASK) APIs, and tools, data and support from the Alexa team.
A $500,000 prize will be awarded to the team selected for creating the best socialbot. The second- and third-place teams will receive prizes of $100,000 and $50,000, respectively. Additionally, a $1 million research grant will be awarded to the winning team’s university if their socialbot achieves the grand challenge of conversing coherently and engagingly with humans for 20 minutes with a 4.0 or higher rating.
An inspiring team of students from the University of Washington won last year’s inaugural competition. Mari Ostendorf, the faculty advisor to last year’s winning team and professor of electrical engineering at the University of Washington, says Alexa Prize was a rewarding experience on many levels.
“Not only did the students build Sounding Board from scratch, they learned from tapping into actual Alexa users across the U.S., and brought together the U of W community to share ideas and secure feedback in order to promote success,” Ostendorf said. “I anticipate the Alexa Prize will be just as exciting for this year’s participants, as each university team works collaboratively to create the future of conversational AI.”
The task of creating a coherent and engaging socialbot continues to be one of the ultimate artificial intelligence challenges. Teams must incorporate and advance several areas of conversational AI within their socialbots, including knowledge acquisition, natural language understanding, natural language generation, context modeling, commonsense reasoning and dialog planning. We will provide students with data, tools, and technical support to help them tackle these problems at scale, along with live interactions and feedback from Alexa’s large user base to help them test ideas and iterate their algorithms much faster than previously possible.
A primary goal of the Alexa Prize is to advance the science of Conversational AI. To this end, the Alexa Prize Proceedings publishes each team’s research resulting from their pursuit of the challenge. Papers from 2017 participants are available here.
The 2018 class of Alexa Prize socialbots will debut this spring. To keep apprised of Alexa Prize developments you can follow the #AlexaPrize hashtag and bookmark our site for updates. In the meantime, you can continue chatting with the winning socialbots from last year’s challenge by simply saying, “Alexa, Let’s chat.” To date, customers have logged more than 100,000 hours of conversation with the 2017 Alexa Prize socialbots.
As we say at Amazon, this is Day 1 for our 2018 Alexa Prize participants. It’s exciting to think that the contributions of this year’s participants will advance us closer to the vision of the Star Trek computer that originally inspired us to create Alexa. Please join me in congratulating our 2018 Alexa Prize participants and in wishing them good luck.
Ashwin Ram is senior manager, AI Science, Alexa Machine Learning, and leads the Alexa Prize. You can follow Ashwin on Twitter @ashwinram