Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces JuRe, a minimal denoising network for time series anomaly detection that challenges the notion that complex architectures are necessary for effective performance. By employing a single depthwise-separable convolutional residual block and a parameter-free structural discrepancy function, JuRe achieves competitive results on multiple benchmarks, ranking second in both the TSB-AD multivariate benchmark and the UCR univariate archive. The findings underscore that the training objective, particularly the denoising approach, is more critical to detection quality than the model's architectural complexity.
JuRe reveals that a simple architecture can outperform complex models in time series anomaly detection by focusing on the right training objectives.
We introduce JuRe (Just Repair), a minimal denoising network for time series anomaly detection that exposes a central finding: architectural complexity is unnecessary when the training objective correctly implements the manifold-projection principle. JuRe consists of a single depthwise-separable convolutional residual block with hidden dimension 128, trained to repair corrupted time series windows and scored at inference by a fixed, parameter-free structural discrepancy function. Despite using no attention, no latent variable, and no adversarial component, JuRe ranks second on the TSB-AD multivariate benchmark (AUC-PR 0.404, 180 series, 17 datasets) and second on the UCR univariate archive by AUC-PR (0.198, 250 series), leading all neural baselines on AUC-PR and VUS-PR. Component ablation on TSB-AD identifies training-time corruption as the dominant factor (ΔAUC-PR = 0.047 on removal), confirming that the denoising objective, not network capacity, drives detection quality. Pairwise Wilcoxon signed-rank tests establish statistical significance against 21 of 25 baselines on TSB-AD. Code is available at the URL https://github.com/iis-esslingen/JuRe.