Karen Jensen

About

During the last decade of the 20th century, Karen served as senior researcher and manager of the Natural Language Processing (NLP) group in Microsoft Research. This group–the first one formed in MSR–aimed to develop a system for computer understanding of human languages, in unrestricted domains, for seven languages: Chinese, English, French, German, Japanese, Korean, and Spanish.

During this period, Karen divided her time between technical and managerial work. Her technical focus was on the development and improvement of a computational syntactic grammar of English. This grammar formed the syntactic component of the initial MS-NLP English system, and was used to jump-start similar components for French, German, and Spanish. Karen worked to initiate techniques for very broad-coverage processing of natural languages, with the goal of making personal computers genuinely able to interact with people in ways that are intelligent, flexible, and easiest for humans. Applications of these techniques began early to move into Microsoft products – for example, in the Word ’97 grammar checker.

The goal remains the same, but in the last couple of decades, NLP research and development techniques have changed dramatically. Compared to 20th-century work, today’s systems handle exponentially more words of text or speech, and instead of using (roughly speaking) phrase-structure rules, they are now powered by neural networks. Karen retired from Microsoft in 2002, but has recently happily returned as a consultant to Microsoft Research  group currently working on NLP.

Karen received her PhD degree from the University of Hawaii (opens in new tab), focusing her doctoral work on pidgin and creole languages, and on the syntax of aspect. She has been a contributor to the computational linguistics community for years.