Alexa Prize 4

DREAM team selected for Alexa Prize Socialbot Grand Challenge 4

On Nov. 3, Amazon published the selection results for the Alexa Prize Socialbot Grand Challenge 4. The Alexa Prize committee selected nine finalists, including a student team from the MIPT Neural Networks and Deep Learning Lab (also known as

The MIPT team is called DREAM. The students are participating for the second year in a row under the guidance of Mikhail Burtsev, head of lab. The team consists of MIPT PhD students also involved in the DeepPavlov project at the Neural Networks and Deep Learning LabDenis KuznetsovDmitry EvseevAnton PeganovAlsu SagirovaDmitry Karpov, and Dilyara Baimurzina, the team leader.

To learn more about the DREAM socialbot architecture created for the previous competition, check out the technical report of the DREAM team from the Alexa Prize 2019 competition. Since the Alexa Prize Socialbot Grand Challenge 3, the DeepPavlov team has transformed the original bot and released the first version of their artificial intelligence-driven ​​assistant — DREAM AI Assistant Demo. This bot is available on the demo website and in the Telegram messenger, in English for now. DREAM AI Assistant Demo blends together almost 40 different chit-chat and task-oriented skills to engage in open domain conversations. It relies on a selection of modern NLP models and components including 14 annotators, 4 post-annotators, and knowledge graph integration. If you have ideas and suggestions for improving the system, then we welcome your feedback.

“This year we have a lot of work ahead of us,” says MIPT PhD student and DREAM’s leader Dilyara Baimurzina. “Over the past year, my team and I have been preparing the engineering infrastructure, as well as the interaction with Amazon services. We have been learning how to test the bot and, most importantly, work in a team. So now we have everything to get right down to implementing our ideas — and reach the finals, of course!”

The Alexa Prize Socialbot Grand Challenge is a competition for undergraduate, graduate, and doctoral students. It was first launched in 2016. The challenge is dedicated to the advancement of conversational artificial intelligence technologies. Teams have to develop a social bot that can communicate fluently with people about news and popular topics such as entertainment, sports, politics, technology, and fashion through Alexa-enabled devices. The ultimate goal is to score 4 out of 5 points and engage the judges in dialogues that last more than 20 minutes in the final round. Bot prototyping will begin in November, and the winner will be announced in August 2021.

The competition is being held for the fourth year, and this time there are nine teams participating, including the past winners of Alexa Prize Socialbot Grand Challenge 3 from Emory University, Stanford, and Czech Technical University. Each team will receive a $250,000 research grant, access to cloud computing in Amazon Web Services, and support from Alexa developers.