ChatPPG.

Wanting Mao

University of Illinois Urbana-Champaign, Urbana, Illinois, USA
+1 more
  • University of Illinois Urbana-Champaign / Google (per-author mapping not given)
ORCID iD

Wanting Mao has 2 papers in the ChatPPG research atlas (2025–2026), linked to 2 models, with work using 9 datasets and 5 co-authors in the atlas. Main topics: accelerometry-FM, MAE, wrist, NHANES, ICLR-2027-sub, multi-rate.

Papers

  • Variable Input Masked Autoencoders for Wrist-Worn Accelerometry (VarMAE)
    2026 · ICLR 2027 submission

    VarMAE learns a single representation across sampling rates and temporal contexts: random crop + downsample creates complementary bottlenecks, and a cross-resolution masked reconstruction loss ties the views. Pretrained on NHANES (~1 week of wrist accelerometry from ~16k subjects); a single encoder matches or outperforms controlled FM baselines (each pretrained on NHANES in the same setup) on eight HAR datasets.

    • Venue: ICLR 2027 submission (under review); arXiv 2610.17383 (2026-10-12)
    • Signals: ACC
    • Form factor: wrist
  • Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings
    2025 · Proceedings of the ACM on Interactive, Mob

    Pulse-PPG is an open-source PPG foundation model trained only on raw PPG from a 100-day field study with 120 participants. Earlier open PPG models were trained on clinical data, and field-trained ones were closed. Pulse-PPG generalized better than clinically trained models across clinical and mobile health tasks in both lab and field settings.

    • Published: Proceedings of the ACM IMWUT, 2025
    • Pretraining: raw field PPG, 100-day study
    • Cohort: 120 participants

Models

  • Pulse-PPG
  • VarMAE: variable-input masked autoencoder for wrist-worn accelerometry

Datasets used in this work

Co-authors

James M. RehgMaxwell A. XuMithun SahaSameer NeupaneSantosh Kumar