OpenMHC: Accelerating the Science of Wearable Foundation Models
cs.LG, cs.AI
Submitted: 2026-06-25
Updated: 2026-11-17
Journal ref: NeurIPS 2026
Code: https://github.com/AshleyLab/OpenMHC
Project page: https://myheartcounts.stanford.edu/openmhc
License: http://creativecommons.org/licenses/by/4.0/
The gist: Mobile and wearable devices offer an unprecedented opportunity for continuous, passive health monitoring and active health coaching.
Terminology
Abstract
Mobile and wearable devices offer an unprecedented opportunity for continuous, passive health monitoring and active health coaching. However, the largest wearable datasets are proprietary, and leading wearable foundation models trained on them are rarely open weight or include reproducible training code. To accelerate open science in wearable health, we introduce OpenMyHeartCounts (OpenMHC), the largest and most comprehensive wearable health dataset broadly accessible to qualified researchers, alongside open-source implementations and weights for recent wearable foundation models. Derived from over a decade of data collected through the My Heart Counts study app, OpenMHC comprises over 67M participant-hours of wearable recordings across up to 19 sensor channels (e.g., step count, heart rate, sleep, workouts) and up to 170 health, lifestyle, mood, and behavioral variables from 11,894 participants. Furthermore, we establish a benchmark for standardized evaluation across three tasks: (1) health and behavioral outcome prediction, (2) multivariate data imputation, and (3) time-series forecasting. Benchmarking classical methods alongside recent wearable and time-series foundation models, we find that while foundation models perform strongly, carefully tuned classical baselines like XGBoost remain highly competitive. By uniting broad-access wearable data with open code and model weights, OpenMHC provides a framework needed to drive reproducible advances in wearable health AI.
Sources
- Wearable Accelerometer Foundation Models for Health via Knowledge Distillation
- Chronos-2: From Univariate to Universal Forecasting
- Toto: Time Series Optimized Transformer for Observability
- This Time is Different: An Observability Perspective on Time Series Foundation Models
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Towards a Personal Health Large Language Model
- Harnessing Vision-Language Models for Time Series Anomaly Detection
- The Anatomy of a Personal Health Agent
- OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data
- MLLM4TS: Leveraging Vision and Multimodal Language Models for General Time-Series Analysis
- MixLinear: Extreme Low Resource Multivariate Time Series Forecasting with 0.1K Parameters
- Towards a General Intelligence and Interface for Wearable Health Data
- Spreading vectors for similarity search
- fev-bench: A Realistic Benchmark for Time Series Forecasting
- LSM-2: Learning from Incomplete Wearable Sensor Data
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis