(personal-health) (wearable-data) (fitness) (sleep) (personalised-insights)


Researchers have introduced PH-LLM, a new large language model (LLM) fine-tuned specifically for understanding personal health data from wearable devices like Fitbits and Pixel Watches. PH-LLM focuses on sleep and fitness data, offering personalised insights and recommendations. The model interprets time-series data to provide customised advice, making it a valuable tool for improving individual health. The paper highlights PH-LLM’s strong performance in understanding and providing recommendations, surpassing human experts in some cases. Additionally, the study developed new datasets and evaluation methods to test and improve the model’s effectiveness.

Business Opportunities

  • Personal Health Applications: Develop apps that utilize PH-LLM to provide users with personalized health insights and recommendations based on their wearable data. This can help users improve their sleep and fitness routines through customised advice.
  • Wearable Device Integration: Partner with manufacturers of wearable devices to integrate PH-LLM into their products, enhancing the value of these devices with advanced, AI-driven health analysis.
  • Fitness and Wellness Coaching: Offer AI-powered fitness and wellness coaching services that leverage PH-LLM’s ability to interpret data and provide tailored advice, potentially creating new revenue streams in the health and wellness industry.
  • Healthcare Providers: Collaborate with healthcare providers to use PH-LLM for monitoring and improving patient health. This can be particularly useful for chronic conditions where continuous monitoring and personalised recommendations are crucial.
  • Telehealth Platforms: Integrate PH-LLM into telehealth platforms to enhance remote patient monitoring and provide actionable health insights, improving patient outcomes and engagement.
  • Health Data Analysis Services: Provide services that analyse large datasets from wearables using PH-LLM, offering detailed reports and insights for research institutions or health-focused organisations.



@leigh.s something up your alleyway

1 Like

Thanks Alex, this is very timely!ucb appreciated :grinning:

1 Like

A follow up to this topic, two examples of companies doing this or about to do this