arXiv AI
This survey reviews deep learning methods for detecting anomalies in railway systems, organizing them by taxonomy of anomaly location, data representation, sensing modality, and temporal traits. It categorizes approaches—convolutional, recurrent, attention-based, autoencoders, GANs, transformers—into classification, prediction, reconstruction, and hybrid paradigms, and discusses data challenges, evaluation, metrics, and deployment issues such as edge‑cloud architectures and hardware constraints. The paper also offers a decision‑oriented framework linking anomaly characteristics, data properties, and operational constraints to guide the selection and deployment of suitable detection solutions.
The paper introduces the Temporal Cycle-Aware Attention Autoencoder with Cross-Signal Consistency (TCAA‑CS) for detecting anomalies in railway passenger door operations. It treats each full opening‑dwell‑closing cycle as a monitoring unit and trains on nominal cycles using a dual‑stream encoder for physical measurements and logical states, an LSTM with temporal attention, and a hybrid anomaly score that fuses reconstruction error, latent‑space deviation, and phase‑aware cross‑signal consistency. On real industrial data, TCAA‑CS achieves 93.8% recall, 97.3% precision, and a 0.5% false‑alarm rate, outperforming other unsupervised baselines and demonstrating real‑time feasibility on an NVIDIA Jetson AGX Xavier.
By Ammar Bouketta, Smail Niar, Hamza Ouarnoughi, Eva Mutuzo Brindle
The paper introduces a deep positive‑unlabeled anomaly detection framework that combines positive‑unlabeled learning with deep models such as autoencoders and deep support vector data descriptions. It addresses the issue of contaminated unlabeled data by approximating anomaly scores for normal data using both unlabeled and labeled anomaly samples, allowing training without labeled normal data. The authors provide a theoretical generalization error bound and demonstrate improved detection performance over existing methods on several datasets.
By Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Yuuki Yamanaka
arXiv:2502. 09194v1 Announce Type: cross Abstract: Generative Artificial Intelligence (AI) techniques have become integral part in advancing next generation wireless communication systems by enabling sophisticated data modeling and feature extraction for enhanced network performance.
By Osman Tugay Basaran, Falko Dressler
arXiv:2507. 15584v2 Announce Type: replace Abstract: Despite the continuous proposal of new anomaly detection algorithms and extensive benchmarking efforts, progress seems to stagnate, with only minor performance differences between established baselines and new algorithms.
By Philipp R\"ochner, Simon Kl\"uttermann, Kevin Kammler, Franz Rothlauf, Emmanuel M\"uller, Daniel Schl\"or
The paper introduces a statistical feature augmentation technique that encodes behavioral interaction statistics into the input space for dynamic graph anomaly detection. Experiments on Reddit, Wikipedia, and MOOC datasets across seven models—both continuous-time and discrete-time—show that this augmentation consistently improves detection performance compared to models trained on original embeddings. The enriched input also facilitates fine-grained post-hoc analysis of behavioral importance, linking classical network analysis with deep learning.
By Philipp Schlinge, Jean-Luc Schnipper, Martin Atzmueller
arXiv:2606. 20323v1 Announce Type: new Abstract: Deep Transfer Learning (DTL) allows for the efficient building of Intelligent Fault Diagnosis Systems (IFDS).
By Giancarlo Santamato, Andrea Mattia Garavagno, Massimiliano Solazzi, Antonio Frisoli
arXiv:2602. 01359v3 Announce Type: replace-cross Abstract: Although recent studies on time-series anomaly detection have increasingly adopted ever-larger neural network architectures such as transformers and foundation models, they incur high computational costs and memory usage, making them impractical for real-time and resource-constrained scenarios.
By Jinju Park, Seokho Kang
While deep learning models have achieved state-of-the-art performance in time series anomaly detection, their complex architectures incur substantial inference overhead. Existing methods typically apply a uniform inference strategy across all data points, which is inefficient given that anomalies are inherently scarce and the vast majority of temporal data consists of predictable normal patterns.
arXiv:2607. 12599v1 Announce Type: new Abstract: Time-series anomaly detection is increasingly important in IoT systems, sensor networks, and edge monitoring applications, where models must operate under strict constraints on memory, latency, and power consumption.
By Raheen Junaid Wani, Smruti R. Sarangi
arXiv:2606. 05070v1 Announce Type: new Abstract: Train delay prediction is an important problem for both passengers and railway operators, yet progress in the field remains difficult to assess due to the lack of standardized datasets, prediction targets, and evaluation protocols.
By Cl\'ement Elliker, Mathis Le Bail, Cl\'ement Mantoux, Jesse Read, Sonia Vanier
arXiv:2608. 01885v1 Announce Type: new Abstract: While deep learning models have achieved state-of-the-art performance in time series anomaly detection, their complex architectures incur substantial inference overhead.
By Zemin Chao, Qianhui Xu, Jianhe Cen, Guangzhi Ge, Xiao Chen, Hoangzhi Wang
arXiv:2609.13940v1 Announce Type: new
Abstract: Effective anomaly detection frameworks are a central pillar of the Industry 4.0 paradigm. In this paper, we introduce an Optimal Transport (OT)-based f...
By Abigail Langbridge, Fearghal O'Donncha, James T Rayfield, Bradley Eck