I will reflect on directions with leveraging advances in machine intelligence to develop systems that enable new forms of competent and fluid human-computer collaboration. I will describe key building blocks and larger architectures that harness machine perception, learning, and inference. Then, I will describe the composition of integrative solutions that draw upon multiple competencies and that operate over extended periods of time. Explorations of such integrative machine intelligence frame research on the coordination of multiple components for sensing and reasoning to create higher-level abstractions, functionalities, and architectures in support of more effecting human-computer symbiosis.
Eric Horvitz is a technical fellow and director of the Microsoft Research lab at Redmond. He has pursued principles and applications of machine intelligence, including efforts on perception, learning, and reasoning and on methods that explore how people and machines can work together as a teams to achieve goals. His research and collaborations have led to fielded systems in healthcare, transportation, human-computer interaction, information retrieval, robotics, and aerospace. He has been elected fellow of AAAI, ACM, and the National Academy of Engineering. He has served on the NSF CISE Advisory Committee, DARPA’s Information and Technology Study Group (ISAT), and Computing Community Consortium (CCC). More information can be found at http://research.microsoft.com/~horvitz.
