Online product recommendation improvement earns doctoral student best paper award

Vaidyanath Areyur Shanthakumar

Vaidyanath Areyur Shanthakumar’s research paper proposes "an elegant solution" to a longstanding data sparsity problem in the area of recommender systems research, says his adviser, Dr. Tathagata Mukherjee.

Michael Mercier / UAH

A doctoral student’s research at The University of Alabama in Huntsville (UAH) to improve the application of artificial intelligence to better understand online user product preferences won the best research paper award at the recent virtual Association for Computing Machinery (ACM) Southeast Conference, the oldest ACM conference in the United States.

Vaidyanath Areyur Shanthakumar is advised by Dr. Tathagata Mukherjee, an assistant professor of computer science at UAH, a part of the University of Alabama System. Shanthakumar’s paper, titled "Item based recommendation using matrix-factorization-like embeddings from deep networks," was presented online and won $200 as part of the award.

Shanthakumar’s research focuses on exploring deep learning-based approaches for recommendation systems and natural language processing.

"Recommender systems are artificially intelligent systems that run in the background of online sites like Netflix and Amazon, and capture user tastes to personalize product recommendations," says Shanthakumar. "This paper proposes a novel approach to solve the longstanding data sparsity problem in the area of recommender systems research and sets a new path forward for further research and improvisation."

"Vaidyanath's work provides an elegant solution to this problem by using the idea of ‘deep neural networks,’" says Dr. Mukerjee. "Using his method, it is possible to compute the recommendations on the entire inventory of a company using reasonable computational resources in an acceptable amount of time. He demonstrated the efficacy of the method on real world data of, one of the larger retailers in the U.S."

Because companies using Shanthakumar’s method can offer recommendations on their entire inventory, it has the potential to positively affect their bottom lines, says Dr. Mukherjee. "This breadth of application area provides an insight into the importance of his work."

Shanthakumar has been working with Dr. Mukherjee for over three years, mainly in artificial intelligence research that explores ways to solve a wide range of problems.

"Currently, I am working with Dr. Mukherjee on a research project with the National Institute of Justice to develop an AI-based interactive mobile app that aids the community correction programs for parolees to avoid recidivism," Shanthakumar says. "In the past, I have worked with Dr. Mukherjee on defense projects with the U.S. Air Force Research Labs. In addition, we also work on recommender systems research, which has high value in commercial applications."

Shanthakumar says he had expected his research to receive attention.

"This research has the ability to have a significant practical impact on commercial recommender systems," he says. "So, I am happy that it received the recognition and was awarded. Moving forward, I would like to contribute more significantly towards transforming the world through artificial intelligence. "



Dr. Tathagata Mukherjee

Jim Steele