Distributed collaborative anomalous sound detection by embedding sharing
This addresses the challenge of privacy-preserving anomaly detection for industrial machine monitoring, though it is incremental as it builds on existing embedding and outlier exposure techniques.
The paper tackled the problem of collaborative anomalous sound detection across multiple clients with private data by proposing a method where clients share embeddings from a pre-trained model, aggregated on a server for detection, resulting in an average AUC improvement of 6.8%.
To develop a machine sound monitoring system, a method for detecting anomalous sound is proposed. In this paper, we explore a method for multiple clients to collaboratively learn an anomalous sound detection model while keeping their raw data private from each other. In the context of industrial machine anomalous sound detection, each client possesses data from different machines or different operational states, making it challenging to learn through federated learning or split learning. In our proposed method, each client calculates embeddings using a common pre-trained model developed for sound data classification, and these calculated embeddings are aggregated on the server to perform anomalous sound detection through outlier exposure. Experiments showed that our proposed method improves the AUC of anomalous sound detection by an average of 6.8%.