Description and Discussion on DCASE 2025 Challenge Task 2: First-shot Unsupervised Anomalous Sound Detection for Machine Condition Monitoring

2025-06-11

AI Summary

Cross-machine anomaly sound detection (ASD) in industrial equipment condition monitoring typically requires extensive labeled data from target machines or labor-intensive hyperparameter tuning. Method: This paper introduces a novel *first-sample unsupervised ASD* task, enabling zero-shot, parameter-free deployment for unseen machine types. We propose the first joint paradigm of *first-sample learning* and *domain generalization*, integrating self-supervised representation learning, unsupervised anomaly modeling, and a cross-domain robust scoring mechanism to construct a prior-free, universal detection framework. Contribution/Results: We establish the first standardized benchmark for first-sample ASD, evaluated uniformly across diverse real-world acoustic recordings from previously unseen machine types. Experiments demonstrate substantial improvements in cross-machine generalization, offering a lightweight, scalable paradigm for industrial intelligent diagnostics.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

This paper introduces the task description for the Detection and Classification of Acoustic Scenes and Events (DCASE) 2025 Challenge Task 2, titled"First-shot unsupervised anomalous sound detection (ASD) for machine condition monitoring."Building on the DCASE 2024 Challenge Task 2, this task is structured as a first-shot problem within a domain generalization framework. The primary objective of the first-shot approach is to facilitate the rapid deployment of ASD systems for new machine types without requiring machine-specific hyperparameter tunings. For DCASE 2025 Challenge Task 2, sounds from previously unseen machine types have been collected and provided as the evaluation dataset. Results and analysis of the challenge submissions will be added following the challenge's submission deadline.

Problem

• First-shot unsupervised anomalous sound detection for machines
• Rapid deployment without machine-specific tuning
• Domain generalization for unseen machine types

Innovation

• First-shot unsupervised anomalous sound detection
• Domain generalization framework approach
• No machine-specific hyperparameter tuning needed