Debanjan Ghosh is a research scientist in the Research & Development division at ETS. He received a Ph.D. in information science from Rutgers University in 2018 and an M.S. in computer science from the University of Buffalo in 2006. Prior to coming to ETS, he was a postdoctoral associate in the brain and cognitive science department at MIT. Between his M.S. and Ph.D., he was a senior software engineer at Thomson Reuters, New York City.
Ghosh is primarily interested in analysis of figurative language (e.g., irony and sarcasm), natural language generation, and argument mining. At ETS, he is involved in automatic item generation and argument mining research. He has published several papers in top-tier NLP conferences, such as the Association for Computational Linguistics (ACL), Empirical Methods in NLP, European Chapter of the ACL, and International Association for the Advancement of Artificial Intelligence Conference on Web and Social Media, and journals, including the top-ranked Computational Linguistics journal. He received a Best Paper Award for his paper The Role of Conversation Context for Sarcasm Detection in Online Interactions at the SIGDIAL 2017 conference.