Disentangling Hierarchical Features for Anomalous Sound Detection Under Domain Shift

TL;DR

This paper introduces a Gradient Reversal-based Hierarchical feature Disentanglement (GRHD) method that improves anomalous sound detection under domain shift by separating domain-related and unrelated features using hierarchical structures.

Contribution

The paper proposes a novel GRHD method that disentangles hierarchical features with gradient reversal, enhancing robustness in ASD under domain shift conditions.

Findings

Significant performance improvement on DCASE 2022 dataset.

Effective separation of domain-related and unrelated features.

Robustness against domain shift in anomalous sound detection.

Abstract

Anomalous sound detection (ASD) encounters difficulties with domain shift, where the sounds of machines in target domains differ significantly from those in source domains due to varying operating conditions. Existing methods typically employ domain classifiers to enhance detection performance, but they often overlook the influence of domain-unrelated information. This oversight can hinder the model's ability to clearly distinguish between domains, thereby weakening its capacity to differentiate normal from abnormal sounds. In this paper, we propose a Gradient Reversal-based Hierarchical feature Disentanglement (GRHD) method to address the above challenge. GRHD uses gradient reversal to separate domain-related features from domain-unrelated ones, resulting in more robust feature representations. Additionally, the method employs a hierarchical structure to guide the learning of fine-grained, domain-specific features by leveraging available metadata, such as section IDs and machine sound attributes. Experimental results on the DCASE 2022 Challenge Task 2 dataset demonstrate that the proposed method significantly improves ASD performance under domain shift.