Ten university teams are competing for the Alexa Prize Socialbot Grand Challenge 3, a university competition to advance the state of the art of conversational AI, and two of those are led by Indian Americans.
Bodhisattwa Majumder, a Ph.D. student at the University of San Diego, is leading Team Bernard while Ashwin Paranjpe is leading Stanford University’s Chirpy Cardinal.
Team Bernard consists of researchers from the artificial intelligence group at the department of computer science and engineering. The research and interests of team members cover a broad swathe of artificial intelligence algorithms and domains, including natural language processing, deep learning, data mining, and reinforcement learning.
As a competitor in the Alexa Prize Challenge, the team, per Amazon, “envisions a future where personal, sensible social conversational agents focus on conversational diversity and consistency in personalized dialogue.”
Majumder’s current research focuses on various aspects in natural language generation and conversational ai tasks — personalization, dialog planning and commonsense reasoning. Prior to UCSD, he was a research engineer at Walmart. He graduated summa cum laude from IIT Kharagpur and has co-authored 10-plus publications, four U.S. patents, and one book on NLP and ML applications.
Paranjpe is currently a third year Ph.D. student in the broad area of NLP and deep learning. His research focus has been about incorporating structure into language, specifically language models. Previously, he did his masters at Stanford, working on data mining, link prediction and graph algorithms research, following his graduation from IIT Bombay.
His team Chirpy Cardinal’s vision is to “create a responsive, empathetic and informative socialbot,” and it aims to have a “holistic approach towards achieving a multi-turn, on-topic and engaging conversation by designing our systems based on the principles of mixed-initiative.”
Paranjpe’s team also includes Indian American student Kaushik Ram Sadagopan.
Nine other Indian American students are part of the university teams competing in the third edition of the challenge.
Nikhil Varghese, a second-year master’s student at the University of California, Santa Cruz, is part of Team Athena; Dravyansh Sharma, a Ph.D. candidate at Carnegie Mellon University in the computer science department, is part of Team Tartan; Harshita S., a third-year computer science Ph.D. student at Emory University where he is part of the Information Retrieval lab, is part of Team Emora; and Ishaan Jain, a first-year master’s student in computer science at UC Davis, and Nikhil Wadhwa of UC Davis, who has 8-plus years of experience working with companies like PwC and Deloitte, and received his bachelor’s from IIT Bombay, are part of the university’s team, Gunrock.
Dheeru Dua, a second year Ph.D. student at the University of California, Irvine, who completed her master’s from LTI at Carnegie Mellon University before joining the Ph.D. program at UCI, has made it to her university’s team, Zotbot. Also part of the team as faculty advisor is Sameer Singh, an assistant professor of computer science at UCI.
Rounding out the list of Indian Americans in the challenge are Arushi Jain, Vihang Agarwal and Sagnik Sinha Roy, who are all part of the University of Michigan’s team, Audrey.
Jain is a master’s student in information in the data science track, while Agarwal is a master’s student with research interests in the field of natural language processing, reinforcement learning and computer vision. Roy is a former software engineer pursuing a masters in machine learning with a penchant for natural language processing and computer vision.
The teams will develop a socialbot, an Alexa skill that converses with users on popular societal topics. Participating teams will advance the state-of-the-art in natural language understanding, dialogue and context modeling, and human-like language generation and expression. They will receive continuous feedback on their inventions in real-world settings.
The grand challenge for the Alexa Prize is to create a socialbot that can engage in a fun, high quality conversation on popular societal topics for 20-minutes and achieve an average rating of at least 4.0/5.0.