K.P. (Suba) Subbalakshmi, a professor in the Department of Electrical and Computer Engineering, received a grant of $271,043 from the United States Army Combat Capabilities Development Command for her project titled “Keyphrase extraction using language embeddings.”
In this project, Subbalakshmi will explore artificial intelligence (AI) and machine learning (ML) architectures that use sentence embeddings to automatically generate keyphrases for any document. Keyphrase extraction is an important innovation in natural language processing (NLP) that presents an efficient way to glean insights from huge amounts of unstructured text data and summarize the topics being discussed. It has several downstream applications; for example, growth in scientific research has resulted in millions of published papers, and it is essential to extract a few keyphrases from the text that are diverse enough to capture the breadth of the topics covered in the text, and also insightful enough to help a specialist understand the depth of content.
Sentence embedding is emerging as a very important part of NLP. This project will explore AI and ML architectures that use sentence embeddings to automatically generate keyphrases for any document. The project will also explore different metrics to evaluate the effectiveness of these keyphrases in representing the topics, and then use those metrics to compare keyphrases and rank them in order of importance.
This one-year project will support two graduate students.