Conversational AI

Emory University Students Win Amazon’s Alexa Prize for their AI Chatbot

A team of Emory University students won Amazon’s 2020 Alexa Socialbot Grand Challenge, a worldwide competition to create that most engaging AI chatbot. The team earned $500,000 for their chatbot named Emora. 

The researchers developed Emora as a social companion that can provide comfort and warmth to people interacting with Alexa-enabled devices. Emora can chat about movies, sports, news of the day, or concerns they might have about themselves or their families during the COVID-19 pandemic. 

“Normally people think of a chatbot as being an intelligent assistant, to answer questions or provide customer service,” graduate student Sarah Fillwock told the Emory News Center.   “We designed a more socially oriented chatbot that could actually show interest in an individual user and provide comfort to people if they wanted it.”

Jinho Choi (center), the faculty advisor for the Emory Alexa Prize team, with graduate students James Finch (left), and Sarah Fillwock, the team leader. Source: Emory University.

Launched in 2016, the Alexa prize challenges students to make breakthroughs in the design of chatbots. This is the third year in a row the Emory team has won the Alexa prize in terms 

To develop the tool the team developed a variety of deep learning models to boost the system’s accuracy. The components included an intent-topic classifier, a response ranker, semantic parsing, coreference resolution, and event extraction models. The success rate was measured by user ratings and how long users continued the conversations. 

“From the beginning, our team slogan has been ‘Emora cares for you. We didn’t set out to design technology that would impress people. We aimed to satisfy people,” Jinho Choi, assistant professor in the Department of Computer Science, told Emory News. “We also wanted to support human relations, especially family relations. We had components that encouraged people to talk about their family members and to keep in touch with them as much as possible.”

On the training side, the team used NVIDIA GPUs with PyTorch, TensorFlow, and MXNet. 

“Practically, all our research uses NVIDIA GPUs to develop deep learning-based NLP models,” Choi said.  “We are currently re-engineering our models to make them lighter so we can integrate them into our next version of Emora.” 

A longer-term goal of the team is to develop a chatbot that can help people with depression or mental health challenges. 

“People with depression often try to avoid talking to others and may become socially isolated,” Choi told the university publication. “A chatbot could potentially help them gain confidence and improve their mood so they could better engage with people.” 

The team has published a technical white-paper about their work, Emora: An Inquisitive Social Chatbot Who Cares For You.

Read more>

Discuss (0)

Tags