LLM-Generated Feature Pools for Time Series Anomaly Detection
This research presents a new approach to time series anomaly detection that uses an LLM to generate feature pools, potentially improving performance on specific datasets.
A new method for time series anomaly detection uses a simple statistical pipeline.
It 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.The candidate pool sets the ceiling;
the search over it is second-order.The method generates 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 best entry on the leaderboard.
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.
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LLM-Generated Feature Pools, time series, anomaly detection, arxiv.
Source attribution: curated from arXiv cs.AI (@Youssef Attia El Hili, Malik Tiomoko, Corinne Ancourt).