aiqbench analysis
This paper introduces CAPS, a novel framework for time series anomaly detection that recovers supervision by simulating context-anchored normal-anomalous pairs without requiring target-domain labels. By leveraging counterfactual recombination to construct a continuous semantic space, the method effectively addresses the scarcity of labeled anomalies and the high context-dependency inherent in temporal data. Experimental results across nine datasets demonstrate superior performance compared to existing unsupervised or surrogate-based approaches.
Supervision Recovery for Time Series Anomaly Detection via Context-Anchored Pairing
arXiv:2609.33610v1 Announce Type: new Abstract: Time series anomaly detection (TSAD) remains challenging not only because anomaly labels are scarce, but also because temporal anomalies are highly context-dependent. Existing methods often rely on unsupervised objectives or surrogate abnormal patterns, providing limited supervision for context-dependent normal--anomalous distinctions. We propose Context-Anchored Pair Supervision (CAPS), a supervision-recovery framework for TSAD. CAPS views ideal anomaly supervision as a matched comparison between normal and anomalous outcomes under the same temporal context, and seeks to recover such supervision without target-domain anomaly labels. Using simulated normal--anomalous pairs, CAPS learns structure and anomaly-semantic representations through reconstruction, background consistency, and within-pair counterfactual recombination. The resulting anomaly representations form a continuous semantic space with coarse modes and induce a sampleable multimodal prior. CAPS conditionally realizes sampled semantics as residual-form effects on target reference trajectories. The resulting context-anchored normal--anomalous counterparts provide temporal supervision for discriminative detector learning. Experiments on nine datasets show that CAPS achieves the strongest aggregate performance across all four evaluation metrics among the compared methods, while complementary ablations and transfer analyses support the roles of context anchoring, semantic disentanglement, and conditional realization.
Abstract:Time series anomaly detection (TSAD) remains challenging not only because anomaly labels are scarce, but also because temporal anomalies are highly context-dependent. Existing methods often rely on unsupervised objectives or surrogate abnormal patterns, providing limited supervision for context-dependent normal--anomalous distinctions. We propose Context-Anchored Pair Supervision (CAPS), a supervision-recovery framework for TSAD. CAPS views ideal anomaly supervision as a matched comparison between normal and anomalous outcomes under the same temporal context, and seeks to recover such supervision without target-domain anomaly labels. Using simulated normal--anomalous pairs, CAPS learns structure and anomaly-semantic representations through reconstruction, background consistency, and within-pair counterfactual recombination. The resulting anomaly representations form a continuous semantic space with coarse modes and induce a sampleable multimodal prior. CAPS conditionally realizes sampled semantics as residual-form effects on target reference trajectories. The resulting context-anchored normal--anomalous counterparts provide temporal supervision for discriminative detector learning. Experiments on nine datasets show that CAPS achieves the strongest aggregate performance across all four evaluation metrics among the compared methods, while complementary ablations and transfer analyses support the roles of context anchoring, semantic disentanglement, and conditional realization.
Artificial Intelligence (cs.AI)
arXiv:2609.33610 [cs.AI]
(or arXiv:2609.33610v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.33610 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 272 words
Related coverage
Anlu: Enabling In-Context Time Series Anomaly Detection in Foundation Models via Counterfactual Supervision
arXiv · Oct 7, 2026
CF-JEPA: Mask-free forward prediction with asymmetric encoder utilization for time-series representation learning
arXiv · Jun 8, 2026
Real-Time Intelligence with IBM Time Series Models on Confluent
Hugging Face · Sep 2, 2026
[AINews] not much happened today
Latent Space · Sep 10, 2026