Meta funds AI research at five US universities
Meta is giving $2 million to five US universities, along with access to its models and its researchers, to apply AI to medicine and the sciences, starting with a set of clinical projects.
Meta announced academic partnerships with the California Institute of Technology, The Ohio State University College of Medicine, Louisiana State University, Indiana University and UTHealth Houston. The deal combines what the company calls "superintelligence models" with research funding and direct collaboration with Meta researchers.
The grants total $2 million across the five schools. They will support research in medicine, physics, chemistry, biology, computer science and mathematics. Meta did not say how the money is split among the schools.
The announcement describes three healthcare projects that use Meta's models, glasses and mixed reality headsets:
- Infant tracheostomy monitoring (LSU). Dr. Michael Dunham leads the Eyes-On project. It is building a computer vision system to detect dislodged tracheostomy tubes and signs of distress in infants who cannot cry out. The team is testing Meta's open-weight SAM models for tracking, and the system is designed to run on bedside cameras with edge computing. A proof-of-concept has been published.
- Dementia-aware conversational AI (Indiana). Dr. David Crandall is developing multimodal methods to analyze conversations involving older adults with dementia. The aim is to help conversational agents, including social robots, adapt to individual users.
- Early heart failure risk (Ohio State). Dr. Yuan Xue, Dr. Yuchi Han and Dr. Xia Ning are studying whether Meta's open-weight models can predict risk of heart failure with preserved ejection fraction (HFpEF). The models draw on routine abdominal CT scans, longitudinal medical records and, when available, ECG data, and are designed to cope with missing inputs. HFpEF accounts for roughly half of all heart failure cases.
Meta did not describe the work at Caltech or UTHealth Houston, or give timelines or say which models are included beyond SAM. The company said more details are on its Research Blog.