mlsanityData to 7 Oct 2026

LLM-Generated Feature Pools for Time Series Anomaly Detection

arXiv 2609.21801 ↗First posted 18 Sep 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

We study how far a simple statistical pipeline can go on univariate time series anomaly detection under a strict selection protocol. The method extracts a small pool of statistics over sliding windows, scores each window with a transductive robust (MAD) model, and selects a feature subset per domain on a held-out tuning split. On TSB-AD-U it reaches 0.529 per-series VUS-PR, above the best neural (0.45) and statistical (0.44) entries on the public leaderboard and within 0.06 of the strongest pretrained foundation model, several of which use more supervision than ours. Ablations locate the cause: across three selection strategies and a hindsight oracle the score moves by 0.031, and across the aggregation grid by 0.096, while changing the candidate pool moves it by 0.226. The candidate pool sets the ceiling; the search over it is second-order. We therefore generate a pool per domain by prompting a multimodal LLM with in-context example windows from that domain. The generated pools match the hand-crafted one under matched selection, and the two cover different domains: selecting over their union improves on the generated pool in all twelve generator-seed pairs and lifts the pipeline to 0.588, matching the performance of the best entry on the leaderboard.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 10 of its 24 references that are in the map. First posted on arXiv 18 Sep 2026.

Builds on

10 of 24 references.

References · 10

  1. 2024Chronos: Learning the Language of Time Seriescited by 712
  2. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 680
  3. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 491
  4. 2024MOMENT: A Family of Open Time-series Foundation Modelscited by 453
  5. 2025TSPulse: Tiny Pre-Trained Models with Disentangled Representations for Rapid Time-Series Analysiscited by 11
  6. 2025Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancycited by 9
  7. 2025Time to Embed: Unlocking Foundation Models for Time Series with Channel Descriptionscited by 2
  8. 2025STREAM-VAE: Dual-Path Routing for Slow and Fast Dynamics in Vehicle Telemetry Anomaly Detectioncited by 2
  9. 2026Tree-Structured Synergy of Large Language Models and Bayesian Optimization for Efficient CASHcited by 2
  10. 2026Matrix Profile for Time-Series Anomaly Detection: A Reproducible Open-Source Benchmark on TSB-ADcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Can LLMs Serve As Time Series Anomaly Detectors?
  2. 2024Large Language Models can Deliver Accurate and Interpretable Time Series Anomaly Detection
  3. 2023MadSGM: Multivariate Anomaly Detection with Score-based Generative Models
  4. 2026Learning Unified Representations of Normalcy for Time Series Anomaly Detection
  5. 2024Can LLMs Understand Time Series Anomalies?
  6. 2026Detecting Time Series Anomalies Like an Expert: A Multi-Agent LLM Framework with Specialized Analyzers
  7. 2025RATFM: Retrieval-augmented Time Series Foundation Model for Anomaly Detection
  8. 2026CARE: A Cascaded Framework for Efficient and Reliable Time Series Anomaly Detection