Improving Robustness of Spectrogram Classifiers with Neural Stochastic Differential Equations
This work addresses robustness issues for signal classification in critical infrastructure domains, but it is incremental as it adapts an existing method to a specific application.
The paper tackled the problem of spectrogram classifiers being vulnerable to noise in signal processing tasks like smart-grid sensing, and introduced Neural Stochastic Differential Equations to improve robustness, achieving a 15% increase in accuracy under low signal-to-noise ratios.
Signal analysis and classification is fraught with high levels of noise and perturbation. Computer-vision-based deep learning models applied to spectrograms have proven useful in the field of signal classification and detection; however, these methods aren't designed to handle the low signal-to-noise ratios inherent within non-vision signal processing tasks. While they are powerful, they are currently not the method of choice in the inherently noisy and dynamic critical infrastructure domain, such as smart-grid sensing, anomaly detection, and non-intrusive load monitoring.