Nurendra Choudhary was an applied science intern with the Amazon Search Team in Palo Alto, California, last summer where he worked on representation learning of products by leveraging the heterogeneous relations between them.
At The Web Conference 2021 last week, Choudhary, a Ph.D. student in computer science at the Sanghani Center, presented “Self-Supervised Hyperboloid Representations Logical Queries over Knowledge Graphs,” his research with data scientists at Amazon and his advisor Chandan Reddy.
It was Reddy’s research on deep learning methods in information retrieval that drew Choudhary to Virginia Tech. “It aligned well with my previous work in social media analytics and I felt that the Sanghani Center would be a great place to develop my expertise in a broader area,” he said.
Choudhary said that he was right. “I have benefited from being able to discuss my own research with a very diverse set of students working on many different problems and getting multiple diverse perspectives and possible solutions to my problems,” he said.
Choudhary’s primary research interest is representation learning with a focus on natural language processing and E-commerce.
Representation learning forms the foundation of most deep learning architectures, he said, and given the potential of change that an improvement in this area could bring, he was extremely interested in contributing to it.
“We notice a lot of E-commerce platforms being spammed by fake reviews,” said Choudhary. “An important pattern in these reviews is a lack of product detail and relevance. With better product and review representations, we can identify the spammers and provide a better customer experience.”
Choudhary has a bachelor’s degree in computer science and master’s degree in computational linguistics, both from the International Institute of Information Technology, Hyderabad, India.
Projected to graduate in 2023, he would like to pursue a career in industry research.