Google Infuses Fitbit with New Personal Health Generative AI Model Based on Gemini
Google will augment Fitbit smart devices with a new generative AI model that can provide personalized health advice through the Fitbit app. The new Personal Health Large Language Model (LLM) is built on Google’s Gemini family of LLMs and processes user data to create an individual’s AI wellness coach.
Fitbit LLM
Google developed the Personal Health LLM by fine-tuning its Gemini LLMs on anonymized health data from research data. Professionals in the field have then helped collate and check the datasets. Google stated that the model is already exhibiting promising reasoning capabilities when tested on sleep medicine certification exam-like practice tests.
The new LLM will analyze various health signals such as sleep patterns, workout intensity, and other physiological data collected by Fitbit wearable devices and Google Pixel phones. Based on this analysis, the AI can suggest actionable tips and adjustments to help users optimize their routines. For example, if the model detects variations in sleep quality, it may recommend modifying workout intensity accordingly, as seen in the screenshot on the right.
“Fitbit and Google Research are working together to build a Personal Health Large Language Model that can power personalized health and wellness features in the Fitbit mobile app, helping people get even more insights and recommendations from the data from their Fitbit and Pixel devices,” Google vice president of engineering and research Yossi Matias explained in a blog post. “This model is being built on Gemini models and fine-tuned on a de-identified, diverse set of health signals from high-quality research case studies. The studies are being collected and validated in partnership with accredited coaches and wellness experts, enabling the model to exhibit profound reasoning capabilities on physiological and behavioral data.”
The new LLM is borne out of the earlier Med-PaLM 2 model, which the Mayo Clinic helped pilot, and is related to the MedLM foundational generative AI models tailored for the healthcare industry. Google also unveiled a new addition to the MedLM suite, specifically for helping analyze and diagnose conditions based on chest X-rays. Google also boasted that its new Articulate Medical Intelligence Explorer (AMIE ) system is as good or better than human doctors in some diagnostics.
Follow @voicebotaiFollow @erichschwartz
Mayo Clinic Begins Testing Google Med-PaLM 2 Generative AI Model
