Adaptable Non-parametric Approach for Speech-based Symptom Assessment: Isolating Private Medical Data in a Retrieval Datastore

TL;DR

This paper introduces a non-parametric speech-based symptom assessment framework that enhances privacy and adaptability by isolating medical data in a retrieval datastore, using self-supervised learning for feature extraction and similarity-based retrieval.

Contribution

It presents a novel non-parametric approach (NoNPSA) that avoids encoding private data in model parameters, allowing efficient updates and improved privacy in healthcare speech assessment.

Findings

Achieves competitive performance with fine-tuning methods

Enhances privacy by isolating data in a retrieval datastore

Enables efficient data updates and adaptability

Abstract

The automatic assessment of health-related acoustic cues has the potential to improve healthcare accessibility and affordability. Although parametric models are promising, they face challenges in privacy and adaptability. To address these, we propose a NoN-Parametric framework for Speech-based symptom Assessment (NoNPSA). By isolating medical data in a retrieval datastore, NoNPSA avoids encoding private information in model parameters and enables efficient data updates. A self-supervised learning (SSL) model pre-trained on general-purpose datasets extracts features, which are used for similarity-based retrieval. Metadata-aware refinement filters the retrieved data, and associated labels are used to compute an assessment score. Experimental results show that NoNPSA achieves competitive performance compared to fine-tuning SSL-based methods, while enabling greater privacy, update efficiency, and adaptability--showcasing the potential of non-parametric approaches in healthcare.