Foundation models for personal health signals

Add to your list(s) Download to your calendar using vCal

• Dimitris Spathis, Google
• Thursday 05 March 2026, 14:00-15:00
• Computer Lab, LT2 and Online.

If you have a question about this talk, please contact Cecilia Mascolo.

https://cam-ac-uk.zoom.us/j/83137163791?pwd=1a513NTE0rQCP3NLnqUZFIxuoW6IAk.1

The unprecedented success of foundation models has transformed our understanding of artificial intelligence, yet their application to personal health and ubiquitous sensing remains a complex frontier. In this talk, I will share my journey in building AI for health monitoring, starting with early efforts to improve data efficiency, robustness, and fairness through self-supervision. A key milestone in this journey was the development of PaPaGei, the first open foundation model for photoplethysmography (PPG). While PaPaGei demonstrated the utility of pre-training on various biosignal datasets, the rigorous demands of real-world applications, such as accurate activity recognition, heart rate, or other biomarker monitoring in dynamic environments, highlight the need for even greater generalization and scale. These challenges motivate the transition to Large Sensor Models that address these hurdles by scaling up both model size and the diversity of user data. This scaling unlocks emergent benefits that smaller models cannot achieve, positioning such foundation models to become the backbone of any future sensing task.

Bio: Dimitris Spathis is a research scientist at Google and a visiting researcher at the University of Cambridge, where he completed his PhD. He was previously a senior research scientist at Nokia Bell Labs, leading efforts in AI for multimodal health. During his studies, he worked at Microsoft Research, Telefonica, and Ocado, while in 2020 he helped start one of the largest studies in audio AI for health (covid-19-sounds.org). He is particularly interested in ways that foundation models and data from personal devices can be helpful for daily health. He studies various topics in AI including data-efficiency, multimodality, model robustness/fairness, and signal processing. His recent work has been featured in international media outlets such as the New York Times, BBC , CNN, Guardian, Washington Post, Forbes, and Financial Times. He serves on the program committees of top AI conferences such as AAAI , IJCAI, and KDD , as well as the editorial boards of Nature Digital Medicine & IEEE Pervasive Computing. More details: https://dispathis.com/

This talk is part of the Mobile and Wearable Health Seminar Series series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge talks
• Computer Lab, LT2 and Online
• Department of Computer Science and Technology talks and seminars
• Interested Talks
• Mobile and Wearable Health Seminar Series
• School of Technology
• Trust & Technology Initiative - interesting events
• bld31
• yk449

Note that ex-directory lists are not shown.