Learning to Adapt to Domain Shifts with Few-shot Samples in Anomalous Sound Detection
This work addresses domain adaptation in anomaly detection for machine health monitoring, but it is incremental as it builds upon prior classification-based methods and meta-learning techniques.
The paper tackles the problem of anomaly detection under domain shifts, such as changes in machine load or environmental noise, by proposing a framework that adapts to new conditions with few-shot samples, achieving around 10% improvement over baselines and matching the best-performing model on an audio dataset.
Anomaly detection has many important applications, such as monitoring industrial equipment. Despite recent advances in anomaly detection with deep-learning methods, it is unclear how existing solutions would perform under out-of-distribution scenarios, e.g., due to shifts in machine load or environmental noise. Grounded in the application of machine health monitoring, we propose a framework that adapts to new conditions with few-shot samples. Building upon prior work, we adopt a classification-based approach for anomaly detection and show its equivalence to mixture density estimation of the normal samples. We incorporate an episodic training procedure to match the few-shot setting during inference. We define multiple auxiliary classification tasks based on meta-information and leverage gradient-based meta-learning to improve generalization to different shifts. We evaluate our proposed method on a recently-released dataset of audio measurements from different machine types. It improved upon two baselines by around 10% and is on par with best-performing model reported on the dataset.