Manas Gaur spent time in multiple doctoral programs before landing at the University of South Carolina. While doing research at Wright State University his advisor was Amit Sheth, who left Wright State to be the founding director of USC’s Artificial Intelligence Institute in 2019.
That same year, Sheth encouraged his former advisee to transfer to USC to join his lab.
To the layperson, Gaur’s research in knowledge-infused learning and knowledge intensive language understanding can get pretty wonky. But in a nutshell, it’s part of an effort to improve upon data-centric statistical learning.
Artificial intelligence depends on data, but you need humans to annotate it before feeding it into the model.
Currently, when a user engages with a model, they only have access to the data itself. Gaur’s goal is to integrate human knowledge into that model. As a result, users of a search engine, for example, will have access to reasonable and explainable insights.
And there are already real-world applications. Gaur’s translational research has been applied in mental health, autonomous driving, chatbots, cybersecurity, online safety and other areas. He was recently awarded an independent EPSRC-UKRI research grant to support research on mental health at Alan Turing Institute, UK.
He is also collaborating with Prisma Health on the deployment of a conversational agent that knows the right questions to ask on clinical questionnaires and how to precisely formulate those questions for safety and diagnosis before engaging with patients.
But why USC? The opportunities for interdisciplinary research. His research was contingent on his ability to communicate with people working in health sciences and social work and use data they had collected.
“It was a very fascinating experience for me, in terms of the collaborations I was able to develop within a short span at USC,” says Gaur, who received a doctorate in computer science in August. Within a year of arriving in Columbia, he was given the opportunity to present his research at the Annual Computing Conference at the SEC Meeting. He was also a part of large team-based National Science Foundation (NSF) and National Institutes of Health (NIH) grants with professors at USC.
Outside of his research, Gaur has tutored high school, undergraduate and graduate students from different backgrounds and institutions. He credits a fellowship at University of Chicago, where he was an Eric and Wendy Schmidt Data Science for Social Good Fellow in 2017, for his interest in giving back. In 2019, he received an AI for Social Good Fellowship from Dataminr Inc.
The interest carried over after he joined Sheth’s lab at USC. He reached out to International Student Services and was directed to SCADR — a program that promotes diversity and inclusion. Through that program, he was able to reach students of different backgrounds and experience levels. He also began mentoring undergraduate students who do not have research experience.
Gaur will continue to mentor students, now as a tenure track assistant professor of computer science at the University of Maryland Baltimore County. Of course, research remains a top priority.
He is excited about an upcoming collaboration with Samsung Research, which is exploring ways to use knowledge-infused AI research in new mobile devices. He also wants to establish an interdisciplinary research center with a foundation in AI and social good and the name KAI2 — a knowledge-infused AI and inference lab at the University of Maryland Baltimore County. He will also be sharing his research at the 2023 Association for the Advancement of Artificial Intelligence New Faculty Highlights Conference in Washington DC.
“I want to establish an NSF/NIH-supported center for research in personalized and trustworthy AI with a targeted focus on digital health, education, online safety and social behavioral sciences,” Gaur says. “It will allow people from diverse backgrounds to engage in research for the benefit of humanity.”