arXiv:2606. 07054v1 Announce Type: cross Abstract: Autonomous LLM agents can pursue hidden malicious objectives through sequences of individually benign actions, making sabotage difficult to detect using standard trajectory-level monitoring.
By Vijitha Mittapalli, Shreyaa Jayant Dani, Satya Srujana Pilli, Snigdha Ansu, Mohammadreza Teymoorianfard, Franck Dernoncourt, Hongjie Chen, Yu Wang, Ryan A. Rossi, Nesreen K. Ahmed
The paper introduces TITAnD, a Trajectory Image Transformer that converts dense and sparse GPS trajectories into a Hyperspectral Trajectory Image (HTI) and applies vision-based classification and segmentation for anomaly detection. It employs a Cyclic Factorized Transformer (CFT) that splits attention along within-day and across-day axes, drastically reducing computational cost and enabling multi-month analysis. Empirical results show TITAnD outperforms existing sparse and dense benchmarks, achieving higher AUC-PR and faster inference than comparable Transformers.
By Md Awsafur Rahman, Chandrakanth Gudavalli, Hardik Prajapati, B. S. Manjunath
The paper introduces VisAnomBench, a new benchmark that augments public time‑series datasets with high‑quality natural‑language anomaly explanations generated by large vision‑language models. Using this benchmark, the authors fine‑tune a parameter‑efficient VLM called VisAnomReasoner, which achieves significantly higher precision and F1 scores than existing baselines on VisAnomBench and also generalizes well to the TSB‑AD‑U benchmark.
By Xiaona Zhou, Muntasir Wahed, Tianjiao Yu, Constantin Brif, Ismini Lourentzou
MAADBench is a refreshable benchmark for anomaly detection in multi‑agent systems powered by large language models. It addresses the challenge of keeping benchmarks current by sampling and coupling generative tasks, generating trace data under configurable LLM backbones, and automatically providing deterministic step‑level labels. The authors evaluated 25 anomaly‑detection methods on 5,200 labeled traces, finding that existing approaches depend heavily on supervision, struggle with subtle MAS‑specific anomalies, and lack robustness across different LLM backbones.
By Lei Ma, Dennis Hofmann, Haowen Xu, Joshua DeOliveira, Peter VanNostrand, Lei Cao, Elke Rundensteiner
arXiv:2605. 22779v2 Announce Type: replace-cross Abstract: Production systems generate millions of log lines daily, yet most anomaly detectors operate at the session or window-level, flagging groups of lines rather than identifying the specific message responsible.
By Huanchi Wang, Zihang Huang, Yifang Tian, Kristina Dzeparoska, Hans-Arno Jacobsen, Alberto Leon-Garcia
arXiv:2609.07941v2 Announce Type: replace
Abstract: In this paper, we propose ReactVAU, a Slow-Fast Decoupled Framework for real-time streaming Video Anomaly Understanding (VAU). Existing VAU methods...
By Chia-Hui Chen, Shih-Ying Yeh, Fu-En Yang, Min-Hung Chen, Shang-Hong Lai