Core AI technologies and applications
At SRI, our researchers work together across disciplines to deliver on the promise of AI and machine learning. We create leading-edge technologies and explore their deployment across a broad range of application areas.
SENSING AND ANALYTICS
Scene recognition and understanding
Person recognition and tracking
Mapping and localization
Text understanding and summarization
REASONING AND PROBLEM SOLVING
Adaptive planning and resource allocation
Decision-making under uncertainty
Deep learning for language, image processing
Learning from demonstration
Hybrid logic/learning architectures
Learning causal models
Funded by a $10.8 million Defense Advanced Research Projects Agency contract, the initiative aims to transform how organizations store and manage data.
SRI International celebrates global recognition of Bioinformatics Research Group and its BioCyc and Pathway Tools collection
SRI’s BioCyc and Pathway Tools publication contributors named to 2022 Highly Cited Researchers List; EcoCyc selected as Global Core Biodata Resource
A powerful legacy; a visionary future
SRI’s Artificial Intelligence Center has driven seminal breakthroughs in AI since the earliest days of the field, including the first mobile-robot (Shakey) and the widely used A* and RANSAC algorithms. Our researchers build on this legacy as we continue to push the boundaries of AI in autonomy, bioinformatics, personal assistants and machine learning.
We present an approach for autonomous sensor control for information gathering under partially observable, dynamic and sparsely sampled environments.
Our new framework provides various measures of RL agent competence stemming from interestingness analysis and is applicable to a wide range of RL algorithms.
We simulate the process of corpus review and word list creation, showing that several simple interventions greatly improve recall as a function of simulated labor.
We present a framework for learning comprehensible models of sequential decision tasks in which agent strategies are characterized using temporal logic formulas.
We describe a large-scale experiment in which non-artificial intelligence subject matter experts (SMEs)—with neither artificial intelligence background nor extensive training in the task—author knowledge bases (KBs) following a challenge problem specification with a strong question-answering component.
Outcome-Guided Counterfactuals for Reinforcement Learning Agents from a Jointly Trained Generative Latent Space
We present a novel generative method for producing unseen and plausible counterfactual examples for reinforcement learning (RL) agents based upon outcome variables that characterize agent behavior.
“SRI’s got that rare mix of people, creativity, technical diversity and excellence, and flexibility to pursue longer-range goals. Of course there are no guarantees, but SRI provides the tools for a motivated person to make things happen.”
Senior Computer Scientist