analiza aiqbench
This research introduces TimeRLM, a recursive language model architecture that addresses context window limitations in time-series analysis by externalizing memory and utilizing code-based signal manipulation. By leveraging reinforcement learning post-training on synthetic benchmarks like AnomalyXL, the system achieves superior anomaly localization performance even when evaluated against real-world clinical and industrial data. The approach demonstrates how integrating reasoning agents with long-context retrieval can solve critical monitoring tasks where traditional single-pass models fail to retrieve brief evidence within vast datasets.
TimeRLM: Recursive Language Models Enable Precise Anomaly Localization in Long-Context Time-Series
arXiv:2608.03391v1 Announce Type: new Abstract: Precise anomaly localization over long-context time series is a crucial task in monitoring applications across clinical care, industrial operations, financial services, and logistics, where brief evidence may hide inside long spans of high-frequency data. Time-Series Language Models (TSLMs) are able to ingest time series data and verbalize findings on anomalies in natural language; however, recent benchmarks report a decrease in retrieval performance at long contexts, mirroring failure modes in text, vision, and audio. In the text domain, Recursive Language Models (RLMs) can recover much of this lost performance by keeping context external to the large language model (LLM), allowing the model to query it through code. We present TimeRLM, an RLM formulation for time-series that sequentially manipulates the signal using code and vision capabilities. We further introduce AnomalyXL, a synthetic long-context anomaly localization benchmark with programmatically injected anomalies that require precise retrieval. We implement five different task categories and two variants: AnomalyXL-MCQ and AnomalyXL-Localize. TimeRLM outperforms every evaluated TSLM and single-pass baseline on four of the five AnomalyXL-Localize tasks, reaching 0.682 IoU on localization and 0.745 on classify-with-evidence, versus at most 0.329 and 0.072 across all baselines. We post-train TimeRLM using reinforcement learning. The resulting model further improves performance and requires approximately one-third as many agent interaction turns as its untrained base model to produce a final answer. On unseen real-world ECG, sleep and software observability recordings, the post-trained TimeRLM retains or improves performance, surpassing TSLMs despite being trained exclusively on synthetic data. Our findings suggest recursive interaction with time-series is an effective approach for long-horizon retrieval.
Authors:Nicolas Zumarraga, Lorenzo Steno, Ning Wang, Max Rosenblattl, Thomas Kaar, Maxwell A. Xu, Kevin O'Sullivan, Markus Kreft, Elgar Fleisch, Paul Schmiedmayer, Patrick Langer, Robert Jakob
Abstract:Precise anomaly localization over long-context time series is a crucial task in monitoring applications across clinical care, industrial operations, financial services, and logistics, where brief evidence may hide inside long spans of high-frequency data. Time-Series Language Models (TSLMs) are able to ingest time series data and verbalize findings on anomalies in natural language; however, recent benchmarks report a decrease in retrieval performance at long contexts, mirroring failure modes in text, vision, and audio. In the text domain, Recursive Language Models (RLMs) can recover much of this lost performance by keeping context external to the large language model (LLM), allowing the model to query it through code. We present TimeRLM, an RLM formulation for time-series that sequentially manipulates the signal using code and vision capabilities. We further introduce AnomalyXL, a synthetic long-context anomaly localization benchmark with programmatically injected anomalies that require precise retrieval. We implement five different task categories and two variants: AnomalyXL-MCQ and AnomalyXL-Localize. TimeRLM outperforms every evaluated TSLM and single-pass baseline on four of the five AnomalyXL-Localize tasks, reaching 0.682 IoU on localization and 0.745 on classify-with-evidence, versus at most 0.329 and 0.072 across all baselines. We post-train TimeRLM using reinforcement learning. The resulting model further improves performance and requires approximately one-third as many agent interaction turns as its untrained base model to produce a final answer. On unseen real-world ECG, sleep and software observability recordings, the post-trained TimeRLM retains or improves performance, surpassing TSLMs despite being trained exclusively on synthetic data. Our findings suggest recursive interaction with time-series is an effective approach for long-horizon retrieval.
Machine Learning (cs.LG)
arXiv:2608.03391 [cs.LG]
(or arXiv:2608.03391v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.03391 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 300 cuvinte
Subiecte conexe
Vertical Advantage: Transforming Industries with Lakebase and Agentic AI
Databricks · 27 aug. 2026
OPD-IAD: From Language Judgment to Industrial Anomaly Detection via On-Policy Self-Distillation
arXiv · 23 iul. 2026
[AINews] High Taste Testers Ep. 1
Latent Space · 17 oct. 2026
Generate Autonomous Business Insights with AI Agent and MCP Servers
AWS · 29 iul. 2026