Benjamin Van Durme is an associate professor of computer science with a secondary appointment in the Krieger School’s Department of Cognitive Science. He is a member of the Center for Language and Speech Processing and leads natural language understanding research at the Human Language Technology Center of Excellence. He is also a research manager at Microsoft, where his team pursues projects in conversational artificial intelligence.
Van Durme’s work focuses on helping people work with large amounts of information; this involves understanding what information is present in documents and images, helping people find that information, and being able to help answer questions about that content. With his collaborators, he publishes on topics in natural language processing, data mining, social media analysis, machine learning, linguistic semantics, and more generally in the fields of AI and cognitive science. His work with colleagues in decompositional semantics is organized at Decomp.io.
Van Durme received his PhD in computer science and linguistics in 2010 from the University of Rochester, where he earlier earned an MS in computer science and a BS and BA in computer science and cognitive science, respectively. He additionally earned an MS in language technologies in 2004 from Carnegie Mellon University.
-
The Conference on Language Modeling is an academic venue with the goal of creating a community of researchers focused on understanding, improving, and critiquing the development of language modeling technology.
-
Johns Hopkins researchers compare GPT models to “sloppy paralegals.”
-
The importance of ambiguity
CategoriesJohns Hopkins researchers find that large language models can handle multiple interpretations of the same sentence—but only if they’re told to.