Neuro-symbolic Neurodegenerative Disease Modeling as Probabilistic Programmed Deep Kernels

TL;DR

This paper introduces a probabilistic deep kernel learning method that combines neural and symbolic approaches for personalized neurodegenerative disease prediction, demonstrating superior accuracy and interpretability over traditional deep learning methods.

Contribution

It proposes a novel Bayesian neuro-symbolic framework that models disease progression without requiring clinical labels, enhancing interpretability and data efficiency.

Findings

Outperforms deep learning in Alzheimer's prediction accuracy

Provides timely and interpretable disease progression models

Leverages domain knowledge through probabilistic programming

Abstract

We present a probabilistic programmed deep kernel learning approach to personalized, predictive modeling of neurodegenerative diseases. Our analysis considers a spectrum of neural and symbolic machine learning approaches, which we assess for predictive performance and important medical AI properties such as interpretability, uncertainty reasoning, data-efficiency, and leveraging domain knowledge. Our Bayesian approach combines the flexibility of Gaussian processes with the structural power of neural networks to model biomarker progressions, without needing clinical labels for training. We run evaluations on the problem of Alzheimer's disease prediction, yielding results that surpass deep learning in both accuracy and timeliness of predicting neurodegeneration, and with the practical advantages of Bayesian nonparametrics and probabilistic programming.