TY - GEN
T1 - Improving Robustness of Spectrogram Classifiers with Neural Stochastic Differential Equations
AU - Brogan, Joel
AU - Kotevska, Olivera
AU - Torres, Anibely
AU - Jha, Sumit
AU - Adams, Mark
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - 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. Currently, these models can be brittle, which makes them susceptible to noisy input. This also means they have sub-optimal stability of explanation outputs. Experts and technicians using these models to make decisions in real world scenarios need assurance that a model is performing as it is supposed to. The classification or prediction outputs it generates should be sound and grounded, not likely to change in the presence of shifting noise landscapes. In this work, we explore the idea of Neural Stochastic Differential Equations (NSDE's) to improve the robustness of models trained to classify time series data and the effect of NSDE's on the explainability of outputs. We then test the effectiveness of these approaches by applying them to a non-intrusive load monitoring (NILM) dataset that consists of simulated harmonic signals injected into a real building.
AB - 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. Currently, these models can be brittle, which makes them susceptible to noisy input. This also means they have sub-optimal stability of explanation outputs. Experts and technicians using these models to make decisions in real world scenarios need assurance that a model is performing as it is supposed to. The classification or prediction outputs it generates should be sound and grounded, not likely to change in the presence of shifting noise landscapes. In this work, we explore the idea of Neural Stochastic Differential Equations (NSDE's) to improve the robustness of models trained to classify time series data and the effect of NSDE's on the explainability of outputs. We then test the effectiveness of these approaches by applying them to a non-intrusive load monitoring (NILM) dataset that consists of simulated harmonic signals injected into a real building.
KW - robustness
KW - signal classification
KW - XAI
UR - http://www.scopus.com/inward/record.url?scp=85210557762&partnerID=8YFLogxK
U2 - 10.1109/MLSP58920.2024.10734825
DO - 10.1109/MLSP58920.2024.10734825
M3 - Conference contribution
AN - SCOPUS:85210557762
T3 - IEEE International Workshop on Machine Learning for Signal Processing, MLSP
BT - 34th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2024 - Proceedings
PB - IEEE Computer Society
T2 - 34th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2024
Y2 - 22 September 2024 through 25 September 2024
ER -