innovating in
Artificial Intelligence
Working across disciplines to deliver on AI’s promise
We’ve engaged in AI research for more than 50 years, creating and delivering world-changing advances like the first mobile robot with the ability to perceive and reason, commercial-quality speech recognition, and the Siri virtual personal assistant.
Today, AI technology — and the creation of new AI technology — are core to the research of many groups across SRI. Our researchers collaborate to build systems that sense, model, and learn to create ever more capable applications. Our approach is human-centered, emphasizing system transparency and explainability to address the critical issues of predictability and reliability of AI-based systems. Contact us to learn more.
Innovating in
“We work across disciplines and in close collaboration with our government and commercial customers to create new technologies, while strengthening the guardrails, so we can all truly benefit from AI’s extraordinary promise.”
Real-world impact
-
Vaalia Health: Causal AI for precision healthcare
With healthcare shifting toward at-home solutions, Vaalia Health is building on SRI’s AI innovations to make sure that efficient care is also effective care.
-
Achieving the potential of autonomous drones
Leveraging SRI navigation technology, Entropy Robotics aims to create safe, resilient autonomous platforms for defense, livestock management, disaster recovery, and much more.
-
Can AI help us find valuable minerals?
SRI’s machine learning-based geospatial analytics platform, already adopted by the USGS, is poised to make waves in the mining industry.
Core AI technologies and applications
Bioinformatics databases
BioCyc database collection
A comprehensive website for sharing fundamental information about biochemical pathways and genomes with researchers around the world
EcoCyc: encyclopedia of E. coli genes and metabolism
A bioinformatics database that describes the genome, metabolic network, and regulatory network of Escherichia coli
Recent publications
-
AI as Collaborative Partner: Rethinking Human-AI Teaming for the Real World
Abstract Much work in human-AI teaming today involves collaboration under fairly constrained settings. Humans supervise AI agents, who are relegated to following orders. The division of tasks is relatively superficial, with independent actions executed in parallel or with simple, linear dependencies between them. Communication is rigid and turn-based, with both parties speaking in complete, unambiguous…
-
ToMCAT: Theory-of-Mind for Cooperative Agents in Teams via Multiagent Diffusion Policies
We introduce ToMCAT, a framework that integrates Theory-of-Mind (ToM) reasoning with a multiagent diffusion model to generate team-aware plans conditioned on agents’ goals and inferred teammate characteristics. Experiments in a simulated domain demonstrate that agents with more ToM capabilities outperform ones without ToM and that ToMCAT’s dynamic replanning mechanism significantly improves resource efficiency while maintaining…
-
Collaborative Metabolic Curation of an Emerging Model Marine Bacterium, Alteromonas macleodii ATCC 27126
A community curation workshop by researchers and students resulted in a new manually-curated genome and metabolic pathway database in BioCyc for the model marine bacterium Alteromonas macleodii ATCC 27126. The curation of this bacterium from peer-review literature illustrates the current knowledge of the organism and highlights gaps in knowledge that serve as starting points for…