Machine learning is a major focus, both as a core research area and as an enabler for natural language processing, computer vision, autonomy, and intelligent assistance. We are pioneering methods to enable effective learning when data is scarce, as well as to understand the limits of learned models. Effective human-machine interaction is a second focus, to ensure that humans can readily understand and direct AI systems. Our research leverages symbolic reasoning for areas such as planning and bioinformatics and hybrid neural-symbolic architectures for tasks that require a combination of learning from data and compiled knowledge.

Founded in 1966, SRI’s Artificial Intelligence Center has played a key role in AI research since the field’s early days. The center has been the source of many seminal contributions to AI, spanning mobile robotics (Shakey), search (the A* algorithm), planning (STRIPS), image understanding (the RANSAC algorithm, TerraVision) and information extraction from text (FASTUS).  

The center has an outstanding track record of moving research into the real world. Spin-offs built using our technology include: Siri (acquired by Apple), Tempo AI (acquired by Salesforce), Meta (acquired by the Chan-Zuckerberg Initiative), Summly (acquired by Yahoo), and Social Kinetics (acquired by Redbrick Health). Recently hatched companies include: Vitrina AI, Kasisto, Pulse, and Confidencial. Our BioCyc family of bioinformatics knowledge bases and tools has been licensed by over 11,000 users to support biological research at universities and corporate labs around the world.