Guangkun Nie
+1 more
- National Institute of Health Data Science, PKU; School of Intelligence Science and Technology, PKU
Guangkun Nie has 4 papers in the ChatPPG research atlas (2024–2026), linked to 1 model, with work using 8 datasets and 20 co-authors in the atlas. Main topics: cross-modal, ppg, ecg, foundation-model, contrastive, ecg-guided.
Papers
- AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling2026 · KDD 2026
AnyPPG pretrains a PPG encoder (Net1D/ResNet-style, 512-dim) using a CLIP-style cross-modal contrastive objective against a frozen ECG foundation model (ECGFounder) instead of PPG-only self-supervised learning. Pretraining data is 109,909 hours / 58,796 subjects of synchronized PPG+ECG from five clinical sources (MC-MED, PulseDB [MIMIC+VitalDB derived], MESA, HSP, CFS) — largely ICU and sleep-lab, not wearable. At inference only PPG is needed (ECG branch is pretraining scaffolding).
- Venue: KDD 2026 (ACM SIGKDD V.2)
- Subjects: 58796
- Signals: ppg, ecg
- Real-world user interactions with large language models for explaining home electrocardiogram reports2026 · npj Digital Medicine
Retrospective analysis of 109,017 conversational sessions / 323,470 dialogue turns from 26,419 users of the Wen-Xin-Wu-Yang AI-ECG system over 16 months, where an LLM explains portable/home ECG reports. 60.01% of interactions happened outside working hours and 57.90% of users came from lower-tier cities. Users showed relatively high retention and moderate depth; their needs moved from result interpretation to medical knowledge, action guidance and health management.
- Venue: npj Digital Medicine (published 2026-10-07; open access CC BY-NC-ND 4.0)
- Signals: ECG, text
- Form factor: other
- PPG Heart Age via ECG-Guided Cross-Modal Pretraining (arXiv:2608.01620)2026 · arXiv preprint
This study estimates heart age from smartwatch PPG, using synchronized smartwatch ECG to guide pretraining so that only PPG is needed in use. It draws on three OPPO cohorts in China with 581,804 participants and 7,452,131 recordings. The PPG-only model estimated age with errors of 5.9 years (r = 0.82) and 4.3 years (r = 0.80) in two external cohorts, and a larger heart-age gap was linked to higher pulse wave velocity.
- Published: arXiv preprint, 2026
- Cohort: 581,804 participants, 7,452,131 recordings (3 OPPO cohorts)
- Signals: smartwatch PPG (ECG-guided pretraining)
- A Scoping Review of Deep Learning Methods for Photoplethysmography Data2024 · arXiv preprint
This scoping review covers deep learning methods for PPG data. It finds that deep learning generally beats handcrafted-feature approaches and is more flexible, but notes ongoing problems: few large high-quality datasets, little real-world validation, and concerns about interpretability, scalability and efficiency.
- Published: arXiv preprint, 2024
- Type: scoping review
- Signals: PPG
Models
Datasets used in this work
- Brno University of Technology Smartphone PPG Database (BUT PPG)
- Cleveland Family Study (CFS): NSRR family sleep-apnea PSG
- Gyro-Acc-PPG (BAMI-I, Wonkwang): 3-channel green wrist PPG + accel + gyro, treadmill
- Human Sleep Project (BDSP): multi-site clinical PSG
- PPG-DaLiA: PPG dataset for motion compensation / HR
- WenXinWuYang: HeartVoice handheld single-lead ECG + finger PPG recordings (private; AnyPPG eval)
- MESA Sleep: NSRR polysomnography + wrist actigraphy (Multi-Ethnic Study of Atherosclerosis)
- UCI Cuff-Less Blood Pressure Estimation (UCI 340; Kachuee 2015)