activity
20242026
collaborators

10 papers

cs.LG2026

Self-Supervised Dynamical System Representations for Physiological Time-Series

Yenho Chen, Maxwell A. Xu, James M. Rehg +1

The effectiveness of self-supervised learning (SSL) for physiological time series depends on the ability of a pretraining objective to preserve information about the underlying phy…

cs.AI2026

How Well Do Multimodal Models Reason on ECG Signals?

Maxwell A. Xu, Harish Haresamudram, Catherine W. Liu +11

While multimodal large language models offer a promising solution to the "black box" nature of health AI by generating interpretable reasoning traces, verifying the validity of the…

cs.LG2025

Prototype-based Self-Supervised Multimodal Learning for PPG and Accelerometry Signals

Wanting Mao, Maxwell A Xu, Harish Haresamudram +3

Modeling multi-modal time-series data is critical for capturing system-level dynamics, particularly in biosignals where modalities such as ECG, PPG, EDA, and accelerometry provide…

cs.LG2025

Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

Mithun Saha, Maxwell A. Xu, Wanting Mao +3

Photoplethysmography (PPG)-based foundation models are gaining traction due to the widespread use of PPG in biosignal monitoring and their potential to generalize across diverse he…

cs.LG2025

LSM-2: Learning from Incomplete Wearable Sensor Data

Maxwell A. Xu, Girish Narayanswamy, Kumar Ayush +22

Foundation models, a cornerstone of recent advancements in machine learning, have predominantly thrived on complete and well-structured data. Wearable sensor data frequently suffer…

eess.SP2025

RelCon: Relative Contrastive Learning for a Motion Foundation Model for Wearable Data

Maxwell A. Xu, Jaya Narain, Gregory Darnell +7

We present RelCon, a novel self-supervised Relative Contrastive learning approach for training a motion foundation model from wearable accelerometry sensors. First, a learnable dis…