Papers › Explainable Time Series Anomaly Detection using Masked Latent Generative Modeling

Explainable Time Series Anomaly Detection using Masked Latent Generative Modeling

21 Nov 2023arXiv:2311.12550archive 2025-07-28

Daesoo Lee, Sara Malacarne, Erlend Aune

We present a novel time series anomaly detection method that achieves excellent detection accuracy while offering a superior level of explainability. Our proposed method, TimeVQVAE-AD, leverages masked generative modeling adapted from the cutting-edge time series generation method known as TimeVQVAE. The prior model is trained on the discrete latent space of a time-frequency domain. Notably, the dimensional semantics of the time-frequency domain are preserved in the latent space, enabling us to compute anomaly scores across different frequency bands, which provides a better insight into the detected anomalies. Additionally, the generative nature of the prior model allows for sampling likely normal states for detected anomalies, enhancing the explainability of the detected anomalies through counterfactuals. Our experimental evaluation on the UCR Time Series Anomaly archive demonstrates that TimeVQVAE-AD significantly surpasses the existing methods in terms of detection accuracy and explainability. We provide our implementation on GitHub: https://github.com/ML4ITS/TimeVQVAE-AnomalyDetection.

PaperPDFCode

Code

ml4its/timevqvae-anomalydetection officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Anomaly DetectionTime SeriesTime Series Anomaly DetectionTime Series Generation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Time Series Anomaly Detection UCR Anomaly Archive TimeVQVAE-AD accuracy 0.708 #1 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive Matrix Profile STUMPY accuracy 0.512 #2 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive MDI accuracy 0.47 #3 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive Matrix Profile SCRIMP accuracy 0.416 #4 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive RCF accuracy 0.387 #5 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive IF accuracy 0.376 #6 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive Convolutional AE accuracy 0.352 #7 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive SR-CNN accuracy 0.3 #8 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive USAD accuracy 0.276 #9 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive AE accuracy 0.236 #10 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive LSTM-VAE accuracy 0.198 #11 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive TranAD accuracy 0.19 #12 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive OC-SVM accuracy 0.088 #13 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive Deep SVDD accuracy 0.076 #14 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive DAGMM accuracy 0.061 #15 of 16 Archive leaderboard report
Time Series Anomaly Detection UCR Anomaly Archive TS-TCC-AD accuracy 0.006 #16 of 16 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections