Article
Computer Science, Information Systems
Bingnan Hou, Changsheng Hou, Tongqing Zhou, Zhiping Cai, Fang Liu
Summary: This study introduces an unsupervised two-step method for detecting and characterizing general network anomalies, which involves finding change-points in large-scale RTT time series and analyzing relationships between links to locate entities likely causing anomalies. Experimental results demonstrate that the proposed method outperforms existing solutions in terms of accuracy and efficiency.
IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT
(2021)
Article
Computer Science, Artificial Intelligence
Xiaoyue Tang, Shan Zeng, Fang Yu, Wei Yu, Zhongyin Sheng, Zhen Kang
Summary: Detecting anomalies in large-scale high-dimensional monitoring data is challenging, especially in the Industry 4.0 environment. This paper proposes a Composite Semantic Augmentation Encoder (CSAE) based on self-supervised contrastive learning for quick anomaly detection in industrial applications. CSAE is a non-sequential deep neural network that can recognize high-level semantic anomalies, improving accuracy and efficiency compared to existing machine learning models.
Article
Engineering, Multidisciplinary
Siya Chen, G. Jin, Xinyu Ma
Summary: This study proposes an improved multivariate Transfer Entropy method for anomaly detection, combined with the CF-LSTM model that utilizes causality features for improved prediction accuracy and sensitivity to anomalies. The proposed method shows improved precision, recall, and F1-score compared to other anomaly detection models, demonstrating its effectiveness.
Article
Statistics & Probability
Wesley Lee, Tyler H. McCormick, Joshua Neil, Cole Sodja, Yanran Cui
Summary: A real-time anomaly detection method was developed for directed activity on large, sparse networks using a dynamic logistic model. This method estimates latent nodal attributes and reduces computational complexity to O(E). After running the algorithm on an enterprise network, it successfully identified a red team attack.
Article
Engineering, Electrical & Electronic
Shenglin Zhang, Zhenyu Zhong, Dongwen Li, Qiliang Fan, Yongqian Sun, Man Zhu, Yuzhi Zhang, Dan Pei, Jiyan Sun, Yinlong Liu, Hui Yang, Yongqiang Zou
Summary: This article proposes an unsupervised KPI anomaly detection approach called AnoTransfer, which combines a novel Variational Auto-Encoder (VAE)-based KPI clustering algorithm with an adaptive transfer learning strategy. Extensive evaluation experiments using real-world data show that AnoTransfer reduces the average initialization time by 65.71% and improves training efficiency by 50.62 times without significantly degrading anomaly detection accuracy.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2022)
Article
Chemistry, Analytical
Suttipong Suttapitugsakul, Yasuyuki Matsumoto, Rajindra P. Aryal, Richard D. Cummings
Summary: In this study, the newly available O-glycoprotease IMPa was used to analyze the glycoproteome of the mouse brain. The results showed that the brain O-glycoproteome only partly overlaps with the N-glycoproteome and contains various sialylated O-glycans and the Tn antigen.
ANALYTICAL CHEMISTRY
(2023)
Article
Automation & Control Systems
Vahab Rostampour, Riccardo M. G. Ferrari, Andre M. H. Teixeira, Tamas Keviczky
Summary: This article addresses two limitations in current distributed model based approaches for anomaly detection in large-scale uncertain nonlinear systems: the high conservativeness of detection thresholds and the requirement for different parties to regularly communicate local measurements. To address these issues, a novel set-based threshold and privacy-preserving mechanism are proposed to ensure robustness and privacy.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
(2021)
Article
Computer Science, Interdisciplinary Applications
Yanbing Wu, Zijian Zhao, Xuejiao Pang, Jin Liu
Summary: This study proposed a semi-supervised anomaly detection model for the initial screening of esophageal endoscopic images. By introducing memory modules and clustering operations, this model performs well in handling unlabeled data and new diseases, and its effectiveness and feasibility were validated in experiments.
JOURNAL OF COMPUTATIONAL DESIGN AND ENGINEERING
(2022)
Article
Engineering, Electrical & Electronic
Soroush Omidvar Tehrani, Afshin Shahrestani, Mohammad Hossein Yaghmaee
Summary: This paper presents an electricity theft detection framework for smart grid data using a hybrid approach combining machine learning algorithms, aiming to discover fraudulent activities. By adding new attack forms and generating malicious data variants to address imbalanced dataset classes, it enhances the accuracy of classifiers. The framework also allows for a trade-off between detection rate and false alarms.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Computer Science, Software Engineering
Amnah Aldayri, Waleed Albattah
Summary: Hajj is an annual Islamic event attended by millions of pilgrims worldwide. Managing and detecting abnormal behaviors in large crowds is a significant challenge for the host country. The current solutions can only handle simple abnormal behaviors, so there is a need for a human abnormal behavior detection approach that can deal with large-scale crowd situations. This study proposes a computer vision-based framework that automatically analyzes video sequences and detects abnormal behaviors. The approach achieves promising results in detecting abnormal pilgrims' behavior.
Article
Engineering, Multidisciplinary
Jing Xu, Xue Li, Puming Wang, Xin Jin, Shaowen Yao
Summary: Distributed Denial of Service (DDoS) attacks consume the resources of networks, making them unable to provide normal services. Accurate detection of DDoS attacks is crucial for network security. However, with the growth of the Internet, the complexity and scale of networks have increased, making it difficult for traditional algorithms to accurately identify attack traffic. In this paper, a novel DDoS attack detection framework is proposed, which uses tensor representation, a multi-modal denoising algorithm based on tensor SVD, and an efficient anomaly detection architecture. Experimental results show that the framework achieves a high detection rate of 98.84% and has good scalability, noise-robustness, and detection speed.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Automation & Control Systems
Qian Wan, Liang Gao, Xinyu Li, Long Wen
Summary: Image anomaly detection and segmentation are crucial for automatic product quality inspection in intelligent manufacturing. This article proposes a novel framework, pretrained feature mapping (PFM), for unsupervised image anomaly detection and segmentation. The proposed framework achieves better results compared to state-of-the-art methods and is also superior in terms of computing time.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Automation & Control Systems
Binyi Su, Zhong Zhou, Haiyong Chen
Summary: The study creates a dataset for anomaly detection in polycrystalline solar cells and provides ground truth bounding boxes, offering convenience for experimental comparisons and evaluations.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Thermodynamics
Li Zhang, Huai Su, Enrico Zio, Zhien Zhang, Lixun Chi, Lin Fan, Jing Zhou, Jinjun Zhang
Summary: The study introduces a method to detect anomalies in an integrated energy system and analyze the system's vulnerability. By combining anomaly detection and complex network theory modeling, the approach effectively exposes security vulnerabilities in the system.
ENERGY CONVERSION AND MANAGEMENT
(2021)
Article
Chemistry, Analytical
Jidong Wang, Gaoxing Jing, Wenxuan Huang, Luhua Xin, Jihui Du, Xiaoqing Cai, Ying Xu, Xi Lu, Wenwen Chen
Summary: Rapid HPV screening method using hydrogel loop-mediated isothermal amplification (LAMP) has been developed, which can provide results in less than 30 minutes and detect HPV infection at the single-cell level in a large-scale parallel manner.
ANALYTICAL CHEMISTRY
(2022)
Article
Geosciences, Multidisciplinary
Sebastian Mueller, Carsten Leven, Peter Dietrich, Sabine Attinger, Alraune Zech
Summary: This study introduces a workflow to estimate geostatistical aquifer parameters using the Python package welltestpy. The analysis is based on semi-analytical drawdown solution and type-curve analysis, which enables the inference of log-transmissivity variance and horizontal correlation length. Sensitivity study shows the impact of observation well positions on parameter estimation quality.
Article
Meteorology & Atmospheric Sciences
Ute Weber, Sabine Attinger, Burkard Baschek, Julia Boike, Dietrich Borchardt, Holger Brix, Nicolas Brueggemann, Ingeborg Bussmann, Peter Dietrich, Philipp Fischer, Jens Greinert, Irena Hajnsek, Norbert Kamjunke, Dorit Kerschke, Astrid Kiendler-Scharr, Arne Koertzinger, Christoph Kottmeier, Bruno Merz, Ralf Merz, Martin Riese, Michael Schloter, HaPe Schmid, Joerg-Peter Schnitzler, Torsten Sachs, Claudia Schuetze, Ralf Tillmann, Harry Vereecken, Andreas Wieser, Georg Teutsch
Summary: MOSES is an observation system designed to study the long-term impacts of dynamic events on environmental systems. It aims to capture these events, from their formation to their end, with high spatial and temporal resolution. It is a mobile and modular system to record energy, water, greenhouse gas, and nutrient cycles, especially the interactions between different compartments of the Earth.
BULLETIN OF THE AMERICAN METEOROLOGICAL SOCIETY
(2022)
Article
Geography, Physical
Anne Koehler, Anneli Wanger-O'Neill, Johannes Rabiger-Voellmer, Franz Herzig, Birgit Schneider, Steven Nebel, Ulrike Werban, Marco Pohle, Manuel Kreck, Peter Dietrich, Lukas Werther, Detlef Gronenborn, Stefanie Berg, Christoph Zielhofer
Summary: This study reconstructs the Holocene deposition history of Loosbach valley in Central Europe and explores the potential tipping point of hydrology during the Late Neolithic occupation. It reveals the interrelation between the hydrological changes and the onset of settlement in the valley floor.
QUATERNARY SCIENCE REVIEWS
(2022)
Article
Environmental Sciences
Jonas L. Schaper, Christiane Zarfl, Karin Meinikmann, Eddie W. Banks, Sandra Baron, Olaf A. Cirpka, Joerg Lewandowski
Summary: This study explicitly models the production, transport, and radioactive decay of Rn-222 in alluvial aquifers, showing substantial variability in production rates within the aquifer. Results indicate that differences in Rn-222 production rates impact estimates of river-to-groundwater travel times, highlighting the importance of explicitly simulating Rn-222 transport and using a combination of natural tracers to reduce uncertainty.
WATER RESOURCES RESEARCH
(2022)
Article
Environmental Sciences
Ruth Maier, Carsten Leven, Daniel Strasser, Bernhard Odenwald, Olaf A. Cirpka
Summary: Horizontal layering in sedimentary bodies affects the directional dependence of hydraulic conductivity in aquifers. Pumping tests can be used to estimate hydraulic anisotropy. Tests with a single depth mainly reflect conditions at that depth, while tests with different depths are needed to identify vertical differences.
Article
Geosciences, Multidisciplinary
Alraune Zech, Sabine Attinger, Alberto Bellin, Vladimir Cvetkovic, Gedeon Dagan, Peter Dietrich, Aldo Fiori, Georg Teutsch
Summary: The goal of this study is to recommend dispersivity values for modeling contaminant transport in groundwater based on a comprehensive analysis of field experiments. The study finds that macrodispersivity coefficients are related to the spatial variability of hydraulic conductivity in groundwater. In the absence of experimental data, practitioners often use ad hoc values for macrodispersivities.
Article
Environmental Sciences
Marie-Madeleine Stettler, Marco Dentz, Olaf A. A. Cirpka
Summary: Macrodispersion in heterogeneous formations is caused by spatial variability of the velocity field. Differential advection interacts with diffusion to determine effective dispersion, while pure advection is reversible and diffusion is irreversible. We found that the reversibility of macrodispersion is bigger for ensemble dispersion than for effective dispersion, challenging the use of the latter as a metric of mixing.
WATER RESOURCES RESEARCH
(2023)
Article
Environmental Sciences
C. Strobel, S. Abramov, J. A. Huisman, O. A. Cirpka, A. Mellage
Summary: This study demonstrates the sensitivity of Spectral Induced Polarization (SIP) to microbially-driven changes in electrical charge storage under natural conditions. The results show that SIP can effectively monitor microbial activity and distinguish it from other polarization processes. This approach opens up new possibilities for in situ monitoring of system reactivity.
JOURNAL OF GEOPHYSICAL RESEARCH-BIOGEOSCIENCES
(2023)
Article
Environmental Sciences
Segolene Dega, Peter Dietrich, Martin Schroen, Hendrik Paasche
Summary: This paper examines the impact of training response variable uncertainty on prediction uncertainties by comparing it with probabilistic prediction obtained using quantile regression random forest. The results provide an uncertainty quantification of the impact on the prediction. The approach is illustrated using the example of probabilistic regionalization of soil moisture derived from cosmic-ray neutron sensing measurements, which produces a regional-scale soil moisture map with data uncertainty quantification for the Selke river catchment in eastern Germany.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2023)
Article
Environmental Sciences
Daniel Altdorff, Sascha. E. E. Oswald, Steffen Zacharias, Carmen Zengerle, Peter Dietrich, Hannes Mollenhauer, Sabine Attinger, Martin Schroen
Summary: A novel rail-borne CRNS system was introduced for continuous monitoring of soil water content along a railway track. The system showed consistent spatial SWC patterns and temporal variations, which can support large scale hydrological modeling and detection of environmental risks.
WATER RESOURCES RESEARCH
(2023)
Editorial Material
Environmental Sciences
Mona Morsy, Erik Borg, Peter Dietrich
Article
Environmental Sciences
Andreas Wieser, Andreas Guentner, Peter Dietrich, Jan Handwerker, Dina Khordakova, Uta Koedel, Martin Kohler, Hannes Mollenhauer, Bernhard Muehr, Erik Nixdorf, Marvin Reich, Christian Rolf, Martin Schroen, Claudia Schuetze, Ute Weber
Summary: Heavy Precipitation Events (HPE) occur when massive amounts of water vapor are transported to a limited area, leading to floods that can cause damage. By combining mobile and stationary observing systems, we can capture the various processes involved in HPE formation and flooding, such as atmospheric transport, precipitation patterns, and runoff dynamics.
ENVIRONMENTAL EARTH SCIENCES
(2023)
Article
Environmental Sciences
Jonas Allgeier, Olaf A. A. Cirpka
Summary: Modern physics-based subsurface-flow models often require many parameters and computationally costly simulations. To expedite the calibration process, we propose using surrogate models based on Gaussian Process Regression (GPR), which allows estimation of the posterior parameter distribution using Markov-Chain Monte Carlo (MCMC) simulations. We compared the GPR-based approach to a Neural Posterior Estimation (NPE) scheme and found that the GPR-based MCMC approach reproduced the data better.
WATER RESOURCES RESEARCH
(2023)
Article
Environmental Sciences
Hannes Mollenhauer, Erik Borg, Bringfried Pflug, Bernd Fichtelmann, Thorsten Dahms, Sebastian Lorenz, Olaf Mollenhauer, Angela Lausch, Jan Bumberger, Peter Dietrich
Summary: This paper introduces a mobile wireless ad hoc sensor network (MWSN) concept that automatically records sufficient close-range data to bridge the gap between standardized and available close-range and satellite remote sensing (RS) data. By cross-calibrating the two systems, comparable spectral characteristics of the data sets could be achieved. Additionally, an analysis of the data reveals the influence of spatial and temporal heterogeneity on the data.