Empowering Multimodal Respiratory Sound Classification with Counterfactual Adversarial Debiasing for Out-of-Distribution Robustness

TL;DR

This paper introduces a counterfactual adversarial debiasing framework for multimodal respiratory sound classification to improve robustness against distribution shifts caused by metadata biases.

Contribution

It presents a novel combination of causal graph-based counterfactual debiasing, adversarial training, and counterfactual data augmentation to reduce metadata biases in respiratory sound classification.

Findings

Outperforms baseline models under distribution shifts

Reduces reliance on non-causal metadata attributes

Enhances generalization across clinical sites

Abstract

Multimodal respiratory sound classification offers promise for early pulmonary disease detection by integrating bioacoustic signals with patient metadata. Nevertheless, current approaches remain vulnerable to spurious correlations from attributes such as age, sex, or acquisition device, which hinder their generalization, especially under distribution shifts across clinical sites. To this end, we propose a counterfactual adversarial debiasing framework. First, we employ a causal graph-based counterfactual debiasing methodology to suppress non-causal dependencies from patient metadata. Second, we introduce adversarial debiasing to learn metadata-insensitive representations and reduce metadata-specific biases. Third, we design counterfactual metadata augmentation to mitigate spurious correlations further and strengthen metadata-invariant representations. By doing so, our method consistently outperforms strong baselines in evaluations under both in-distribution and distribution shifts. Code is available at https://github.com/RSC-Toolkit/BTS-CARD.