Paul Bennett

I am a Partner Researcher and manager of the Productivity and Intelligence group in Microsoft Research. I am interested in the development, improvement, and analysis of machine learning methods with a focus on systems that can aid in the automatic analysis of natural language as components of adaptive systems or information retrieval systems. My current focus is on contextually intelligent assistants. I also maintain an active interest in contextual and personalized search, enriched information retrieval, active sampling and learning, hierarchical and large-scale classification, and human computation and preferences.

My past work has examined a variety of areas — primarily ensemble methods, calibrating classifiers, search query classification and characterization, and redundancy and diversity, but also extending to transfer learning, machine translation, recommender systems, and knowledge bases. In addition to my research, I engage in a variety of professional service activities for the machine learning, data mining, and information retrieval communities.

Before coming to Microsoft, I obtained my Ph.D. from the Computer Science Department at Carnegie Mellon University under Jaime Carbonell and John Lafferty. Prior to that I worked with Ray Mooney and Robert Causey during my undergraduate days in the Computer Science, Philosophy, and Plan II Honors departments at the University of Texas at Austin.