Cornell researchers are teaming up with the Toyota Research Institute (TRI) as part of a new phase of the institute's collaborative University Research Program focused on advancing artificial intelligence and robotics, among other research areas.
Research projects spanning 31 universities were announced May 18 as part of the University Research Program 3.0, with each project being co-led by university researchers and a Toyota Research Institute co-investigator working as peers to ensure that fundamental research and real-world application evolve together. Leading one of Cornell's projects are Hadas Kress-Gazit, the Geoffrey S.M. Hedrick Senior Professor, together with Guy Hoffman, associate professor - both in the Sibley School of Mechanical and Aerospace Engineering in the Cornell Duffield College of Engineering. Leading a separate project is Angelina Wang, assistant professor of information science in the Cornell Bowers College of Computing and Information Science and based at Cornell Tech.
Kress-Gazit and Hoffman's project, "Human-Robot Interaction with LBMs: Mitigating Collaborative Task Failures," explores how robots powered by large behavior models can better collaborate with people when tasks go wrong or conditions change unexpectedly. The research aims to detect failures in the human-robot collaboration, visualize them to the person, and repair the interaction.
The work builds on Kress-Gazit and Hoffman's research on human-robot interaction, how people react to robot failures, and techniques for detecting and predicting robot failure.
Wang's project, "Human-AI Co-Evolution: Assessing Risks and Benefits of Personalized Conversational AI Agents," examines how AI chatbots personalize interactions with users and how those systems can be designed responsibly. The project will study how personalization affects trust, comfort and user experience while developing methods to evaluate and improve chatbot behavior.
Wang's research addresses growing questions surrounding societal impacts of AI, including conversational AI systems that are increasingly adapting to individual users across education, work and daily life.
"At TRI, we believe breakthrough research happens through deep collaboration and that we are better together," said Kate Tsui, senior manager of TRI's University Research Program. "These are not passive sponsorships. TRI co-investigators work closely with faculty and students throughout the research lifecycle. By combining academic excellence with real-world challenges, we are building pathways from foundational research to meaningful impact."
