Reconstructive Reservoir Computing to Detect Anomaly in Time-series Signals

Kato, Junya; Tanaka, Gouhei; Nakane, Ryosho; Hirose, Akira

doi:10.36227/techrxiv.17704682.v1

“…Thus, by setting the threshold appropriately for E(n), we can detect anomaly. We also obtained similar results from the experiment using fan data [41].…”

Section: Anomaly Detection By Reconstructionsupporting

confidence: 70%

Reconstructive reservoir computing for anomaly detection in time-series signals

Kato,

Tanaka,

Nakane

et al. 2024

NOLTA

1

0

View full text Add to dashboard Cite

We propose reconstructive reservoir computing (RRC) which performs better anomaly detection in time-series signals than forecasting-based methods. In this paper, reconstruction means that a neural network generates past input signals. RRC reconstructs a past normal signal for anomaly detection using an echo state network which can learn quickly and stably. We expect that it is easier to restore a past normal signal than to predict an unknown future normal signal. For anomaly detection, we compute an instantaneous reconstruction error. The reconstruction error larger than a threshold is a sign of anomaly. We conduct experiments using sound data obtained from a pump. In the experiments, we pay attention to a time lag between input and output to be reconstructed since we assume that an excessive time lag makes reconstruction difficult due to signal attenuation in the network. Experimental results show that if the time lag is moderate, the reconstruction error of the normal signal is lower than the forecasting error of the same signal. Furthermore, we show that RRC with the appropriate time lag has a better anomaly detection performance index than forecasting-based methods.

show abstract

“…In most reservoir-based anomaly detection approaches [23], [24], [25], [26], [27], [28], the model output predicts the future input (i.e., M = L and f out (t) := f in (t + ∆t)). Then, the magnitude of the prediction error, such as the squared prediction error (SPE) ∥f out (t)−z(t)∥ 2 , serves as an anomaly score.…”

Section: B Standard: Prediction Error-based Anomaly Detectionmentioning

confidence: 99%

“…The most widely adopted anomaly score when using a reservoir for time-series anomaly detection is the squared prediction error (SPE) E S (t) [23], [24], [25], [26], [27], [28] defined as…”

Section: B Anomaly Measures For Reservoirsmentioning

confidence: 99%

See 1 more Smart Citation

Mahalanobis Distance of Reservoir States for Online Time-Series Anomaly Detection

Tamura¹,

Fujiwara²,

Aihara³

et al. 2023

Preprint

0

View full text Add to dashboard Cite

<p>Automated anomaly detection in time-series data has important applications in modern society. In practical settings, sequence patterns must often be evaluated in real-time under limited computational resources and scarce training data, presenting significant challenges. Reservoir computing models can help reconcile the trade-off between computational expense and precision in such situations. We propose the Mahalanobis distance of reservoir states (MD-RS) method, as an alternative to standard reservoir-based methods that rely on prediction error. MD-RS can effectively highlight anomalies because it measures the deviation from the subspace that minimally encompasses reservoir responses to normal inputs. Our semi-supervised MD-RS method only requires training data labeled as normal, making it suitable for practical scenarios under limited training data. Both the training and inference can be performed online with reduced computational cost, enabling real-time anomaly detection. We apply the MD-RS method to a benchmark dataset with a short training phase from the UCR time series anomaly archive and demonstrate its superiority in terms of accuracy, reliability, and robustness over prediction error-based methods when employing the same reservoir. Additionally, we propose the response dimension increase (RDI) index for initial hyperparameter tuning, which measures the increase in the effective dimension of the reservoir trajectories. These results position our method as a leading choice for applications requiring rapid and accurate anomaly detection.</p>

show abstract

Mahalanobis Distance of Reservoir States for Online Time-Series Anomaly Detection

Tamura¹,

Fujiwara²,

Aihara³

et al. 2023

Preprint

0

View full text Add to dashboard Cite

<p>Automated anomaly detection in time-series data has important applications in modern society. In practical settings, sequence patterns must often be evaluated in real-time under limited computational resources and scarce training data, presenting significant challenges. Reservoir computing models can help reconcile the trade-off between computational expense and precision in such situations. We propose the Mahalanobis distance of reservoir states (MD-RS) method, as an alternative to standard reservoir-based methods that rely on prediction error. MD-RS can effectively highlight anomalies because it measures the deviation from the subspace that minimally encompasses reservoir responses to normal inputs. Our semi-supervised MD-RS method only requires training data labeled as normal, making it suitable for practical scenarios under limited training data. Both the training and inference can be performed online with reduced computational cost, enabling real-time anomaly detection. We apply the MD-RS method to a benchmark dataset with a short training phase from the UCR time series anomaly archive and demonstrate its superiority in terms of accuracy, reliability, and robustness over prediction error-based methods when employing the same reservoir. Additionally, we propose the response dimension increase (RDI) index for initial hyperparameter tuning, which measures the increase in the effective dimension of the reservoir trajectories. These results position our method as a leading choice for applications requiring rapid and accurate anomaly detection.</p>

show abstract

Reconstructive Reservoir Computing to Detect Anomaly in Time-series Signals

Cited by 3 publications

References 23 publications

Reconstructive reservoir computing for anomaly detection in time-series signals

Reconstructive reservoir computing for anomaly detection in time-series signals

Mahalanobis Distance of Reservoir States for Online Time-Series Anomaly Detection

Mahalanobis Distance of Reservoir States for Online Time-Series Anomaly Detection

Contact Info

Product

Resources

About