Dan Roth (Hebrew: דן רוט), associated with Haifa, Israel, is the Eduardo D. Glandt Distinguished Professor of Computer and Information Science at the University of Pennsylvania and the Chief AI Scientist at Oracle.[5,4] Until June 2024, Roth was a VP and distinguished scientist at AWS AI. In his role at AWS, Roth led over the last three years the scientific effort behind the first-generation Generative AI products from AWS, including Titan Models, Amazon Q efforts, and Bedrock, from inception until they became generally available.
Roth earned his B.A. summa cum laude in mathematics from the Technion, Israel, and received his Ph.D. in computer science from Harvard University in 1995 under Leslie Valiant.[5] He taught at the University of Illinois at Urbana-Champaign from 1998 to 2017 before moving to the University of Pennsylvania, with Ming-Hsuan Yang among his students.
His research focuses on computer science, machine learning, natural language processing, automated reasoning, and information extraction, including joint learning and inference (such as ILP formulations of NLP tasks), machine learning for NLP, and probabilistic reasoning.[1] Roth is an ACM Fellow, a recipient of the IJCAI John McCarthy Award, and maintains his homepage at http://www.cis.upenn.edu/~danroth.[2,3]
Professional career
Roth's research focuses on the computational foundations of intelligent behavior. He develops theories and systems pertaining to intelligent behavior using a unified methodology centered on the idea that learning plays a central role in intelligence. His work centers on the study of machine learning and inference methods to facilitate natural language understanding, pursuing interrelated lines of work spanning fundamental questions in learning and inference to the development of advanced machine learning-based tools for natural language processing applications.[12,11] Roth has made seminal contributions to the fusion of learning and reasoning, machine learning with weak and incidental supervision, and machine learning and inference approaches to natural language understanding.[13,14] He wrote the first paper on zero-shot learning in natural language processing—a 2008 paper by Chang, Ratinov, Roth, and Srikumar published at AAAI'08, where the paradigm was termed dataless classification.[15] He has also worked on probabilistic reasoning (including its complexity and probabilistic lifted inference), Constrained Conditional Models (ILP formulations of NLP problems), constraints-driven learning, part-based (constellation) methods in object recognition, and response-based learning.[12,16,17,18,20,21] Additionally, Roth has developed natural language processing and information extraction tools that are broadly used commercially and by researchers, including tools for NER, coreference resolution, wikification, SRL, and ESL text correction.[12,16,17,18,20,21]
Roth is a Fellow of the American Association for the Advancement of Science (AAAS), the Association for Computing Machinery (ACM), the Association for the Advancement of Artificial Intelligence (AAAI), and the Association of Computational Linguistics (ACL).[9,6,7,8] He is a co-founder of NexLP, Inc., a startup applying natural language processing and machine learning to legal and compliance domains, which was acquired by e-discovery software company Reveal, Inc. in 2020.[22] He currently serves on the scientific advisory board of the Allen Institute for AI.[23]
.jpg)