Quantum machine learning framework for longitudinal biomedical studies
Journal:
arXiv
Published Date:
Apr 24, 2025
Abstract
Longitudinal biomedical studies play a vital role in tracking disease
progression, treatment response, and the emergence of resistance mechanisms,
particularly in complex disorders such as cancer and neurodegenerative
diseases. However, the high dimensionality of biological data, combined with
the limited size of longitudinal cohorts, presents significant challenges for
traditional machine learning approaches. In this work, we explore the potential
of quantum machine learning (QML) for longitudinal biomarker discovery. We
propose a novel modification to the instantaneous quantum polynomial time (IQP)
feature map, designed to encode temporal dependencies across multiple time
points in biomedical datasets. Through numerical simulations on both synthetic
and real-world datasets - including studies on follicular lymphoma and
Alzheimer's disease - we demonstrate that our longitudinal IQP feature map
improves the ability of quantum kernels to capture intra-subject temporal
patterns, offering a promising direction for QML in clinical research.
