The Language & Information Technologies Group at Microsoft Research AI focuses on enabling machines to understand and communicate in natural language, answer complex questions, assist with task completion and understand how people interact with information. Our research interests are at the intersection of natural language understanding, machine learning and information retrieval. Our work involves analyzing large-scale text and user behavior data to understand how users interact with information, build models for extracting knowledge from unstructured text and help users complete tasks and find information efficiently.
Examples of current directions include natural language processing (e.g. question answering, semantic parsing, natural language interfaces, cross-lingual models, etc.), machine learning (specifically deep learning, transfer learning and weakly supervised learning) and information access (e.g. conversational search, recommendation and user modeling).
Our work has application in language understanding, question answering, intelligent assistants, search and personalization. Our work has been shipped in several experiences and products such as Bing, Cortana, Office 365, and Dynamics 365.