Deep Anomalies Research Articles

An Object‐Oriented Bayesian Gravity Inversion Scheme for Inferring Density Anomalies in Planetary Interiors

AbstractGravity inversions have contributed greatly to our knowledge of the interior of planetary bodies and the processes that shaped them. However, previous global gravity inversion methods neglect the inference of mantle density anomalies when using techniques to decrease the non‐uniqueness of the inversion. In this work, we present a novel global gravity inversion algorithm, named THeBOOGIe, suited to inferring global‐scale density anomalies within the crust and mantle of planetary bodies. The algorithm embraces the nonuniqueness inherent in gravity inversions by not prescribing at the outset a density interface or depth range of interest. Instead, the method combines a Bayesian approach with a flexible incorporation of prior geological or geophysical information to infer density anomalies at any depth. A validation test using synthetic lunar‐like gravity data shows that THeBOOGIe can constrain the lateral location of crustal density anomalies but tends to overestimate their thicknesses. Importantly, THeBOOGIe can detect deep mantle density anomalies and quantify the level of confidence in the inferred density models. Our results show that THeBOOGIe can provide complementary information to one‐dimensional seismic models of the interior of the terrestrial planets and the Moon by constraining density anomalies that are not spherically symmetric. Additionally, THeBOOGIe is specially suited to constraining the interior of partially differentiated bodies where these large‐scale density anomalies are more likely to exist. Finally, thanks to the flexible use of priors, THeBOOGIe is an essential tool to understand the interior of planetary bodies lacking additional constraints.

Deep Anomaly Detection and Search via Reinforcement Learning (Student Abstract)

Deep Anomaly Detection and Search via Reinforcement Learning (Student Abstract)

Deep learning-based sustainable subsurface anomaly detection in Barker-coded thermal wave imaging

Deep learning-based sustainable subsurface anomaly detection in Barker-coded thermal wave imaging

Anomaly Detection of Breast Cancer Using Deep Learning.

Cancer is one of the deadliest diseases facing humanity, one of the which is breast cancer, and it can be considered one of the primary causes of death for most women. Early detection and treatment can significantly improve outcomes and reduce the death rate and treatment costs. This article proposes an efficient and accurate deep learning-based anomaly detection framework. The framework aims to recognize breast abnormalities (benign and malignant) by considering normal data. Also, we address the problem of imbalanced data, which can be claimed to be a popular issue in the medical field. The framework consists of two stages: (1) data pre-processing (i.e., image pre-processing); and (2) feature extraction through the adoption of a MobileNetV2 pre-trained model. After that classification step, a single-layer perceptron is used. Two public datasets were used for the evaluation: INbreast and MIAS. The experimental results showed that the proposed framework is efficient and accurate in detecting anomalies (e.g., 81.40% to 97.36% in terms of area under the curve). As per the evaluation results, the proposed framework outperforms recent and relevant works and overcomes their limitations.

MUD enabled deep learning framework for anomaly detection in IoT integrated smart building

Nowadays, many Internet of Things (IoT) devices of different types are used in creating smart applications like smart cities, smart industries, smart environments, and the applications of industry-4.0. IoT devices are used for different purposes, such as security, remote monitoring, resource allocation, threats, ecosystems, and vulnerabilities. This paper proposed a deep learning algorithm-based solution to tighten the security level in the IoT-Smart environment network. The Intrusion Detection System (IDS) considered in this paper is Network IDS, which investigates the manufacturer usage description, digital twins, and deep learning-based user behavior information. IoT devices' communication and the users in smart buildings are automatically connected in the Intelligent Communication system. Since many devices and users are interconnected in smart buildings, the probability of cyber-attack is high. Thus, better security is needed in smart buildings and smart environments. It should focus on securing IoT devices, users, and their communication. Hence, this paper developed a deep learning-based anomaly detection framework to dynamically monitor the issues and problems with MUD profiles and detect the anomaly behavior. The Manufacturer Usage Description (MUD) profiles, dynamic user behavior, IoT devices' traffic data the pattern of abnormal/anomaly traffic at the device level is predicted while traffic occurs. The MUD-ML-based model is implemented in Python software, verifying the results.

Benchmarking framework for anomaly localization: Towards real-world deployment of automated visual inspection

Localizing defects in products is a critical component of industrial pipelines in manufacturing, retail, and many other industries to ensure consistent delivery of high quality products. Automated anomaly localization systems leveraging computer vision have the potential to replace laborious and subjective manual inspection of products. Recently, there have been tremendous efforts in this research domain investigating deep learning-based anomaly localization methods. However, such proposed methods, mainly considering product-specific evaluation, cannot be directly implemented for real-world use. Therefore, despite the advancements, there is still a gap between research and deployment of those methods to real-world production environments. Implementing any automated solution for manufacturing can involve a steep upfront investment. It is important to develop an industry-friendly benchmarking framework to ensure the feasibility and robustness of an automated quality control system. In this paper, we present a new anomaly localization benchmarking framework considering different aspects - (1) understand the performance of models in a generalizable product-agnostic manner, (2) explore pros and cons to find the most optimal modeling approach, (3) develop an efficient training and inference scheme with defect-free training samples and very few defective samples for evaluation, and (4) perform an ablation study of threshold estimation techniques to determine optimal threshold level for segmentation. We release a newly-labeled dataset for the research community with product-agnostic categorization of defective product images. To the best of our knowledge, this is the first anomaly localization work on developing a benchmarking framework focusing on real-world use. We believe domain experts from different industries will find this useful and can gain valuable insights to deploy automated visual inspection in production pipelines.

Deep learning for predicting future lesion emergence in high-risk breast MRI screening: a feasibility study

BackgroundInternational societies have issued guidelines for high-risk breast cancer (BC) screening, recommending contrast-enhanced magnetic resonance imaging (CE-MRI) of the breast as a supplemental diagnostic tool. In our study, we tested the applicability of deep learning-based anomaly detection to identify anomalous changes in negative breast CE-MRI screens associated with future lesion emergence.MethodsIn this prospective study, we trained a generative adversarial network on dynamic CE-MRI of 33 high-risk women who participated in a screening program but did not develop BC. We defined an anomaly score as the deviation of an observed CE-MRI scan from the model of normal breast tissue variability. We evaluated the anomaly score’s association with future lesion emergence on the level of local image patches (104,531 normal patches, 455 patches of future lesion location) and entire CE-MRI exams (21 normal, 20 with future lesion). Associations were analyzed by receiver operating characteristic (ROC) curves on the patch level and logistic regression on the examination level.ResultsThe local anomaly score on image patches was a good predictor for future lesion emergence (area under the ROC curve 0.804). An exam-level summary score was significantly associated with the emergence of lesions at any location at a later time point (p = 0.045).ConclusionsBreast cancer lesions are associated with anomalous appearance changes in breast CE-MRI occurring before the lesion emerges in high-risk women. These early image signatures are detectable and may be a basis for adjusting individual BC risk and personalized screening.Relevance statementAnomalies in screening MRI preceding lesion emergence in women at high-risk of breast cancer may inform individualized screening and intervention strategies.Key points• Breast lesions are associated with preceding anomalies in CE-MRI of high-risk women.• Deep learning-based anomaly detection can help to adjust risk assessment for future lesions.• An appearance anomaly score may be used for adjusting screening interval times.Graphical

Geophysical Mapping by Electromagnetic Induction of Gold Occurrences in Birimian Formations of Liptako: Case of Sorbon Haoussa Sector (Souhwest Niger)

The survey zone corresponds to the Sirba greenstone belt, one of Liptako Birimian Formations (LBF) of Niger. Previous aeromagnetic and geochemical works reveal the significant gold occurrences. The present study integrates geophysical mapping of these occurrences by electromagnetic induction on field. A multi-frequencies system of 10 frequencies was used for depth and subsurface investigation of resistive anomalies. The interpretation of obtained results shown two types of resistive anomalies, corresponding to the conductive units: deep anomalies obtained with low frequencies and subsurface anomalies detected with high frequencies. The subsurface anomalies were detected around the 15 meters and the deep anomalies were recorded beyond 80 meters. All of the detected resistive anomalies are in a lenticulars shapes oriented N-S or E-W. The analysis of combined resistivity map from all frequencies shown three zones of resistive anomaly presenting a high potential in conductive metals: The first zone Z1 (most important), located in Eastern studied zone covering, covering a surface of 3.5 Km2, the second zone Z2 is located in West and covers an area of 2.9 Km2 and the third anomaly zone Z3 is located in South with a surface of 0.96 Km2. On field, these conductive units correspond to auriferous quartz veins and/or highly altered manganese schists, exploited by using the artisanal mining wells, up to 20 meters in depth.

Deep Learning-Based Anomaly Detection in Video Surveillance: A Survey.

Anomaly detection in video surveillance is a highly developed subject that is attracting increased attention from the research community. There is great demand for intelligent systems with the capacity to automatically detect anomalous events in streaming videos. Due to this, a wide variety of approaches have been proposed to build an effective model that would ensure public security. There has been a variety of surveys of anomaly detection, such as of network anomaly detection, financial fraud detection, human behavioral analysis, and many more. Deep learning has been successfully applied to many aspects of computer vision. In particular, the strong growth of generative models means that these are the main techniques used in the proposed methods. This paper aims to provide a comprehensive review of the deep learning-based techniques used in the field of video anomaly detection. Specifically, deep learning-based approaches have been categorized into different methods by their objectives and learning metrics. Additionally, preprocessing and feature engineering techniques are discussed thoroughly for the vision-based domain. This paper also describes the benchmark databases used in training and detecting abnormal human behavior. Finally, the common challenges in video surveillance are discussed, to offer some possible solutions and directions for future research.

Domain Adaptation-Based Automated Detection of Retinal Diseases from Optical Coherence Tomography Images

Purpose To verify the effectiveness of domain adaptation in generalizing a deep learning-based anomaly detection model to unseen optical coherence tomography (OCT) images. Methods Two datasets (source and target, where labelled training data was only available for the source) captured by two different OCT facilities were collected to train the model. We defined the model containing a feature extractor and a classifier as Model One and trained it with only labeled source data. The proposed domain adaptation model was defined as Model Two, which has the same feature extractor and classifier as Model One but has an additional domain critic in the training phase. We trained the Model Two with both the source and target datasets; the feature extractor was trained to extract domain-invariant features while the domain critic learned to capture the domain discrepancy. Finally, a well-trained feature extractor was used to extract domain-invariant features and a classifier was used to detect images with retinal pathologies in the two domains. Results The target data consisted of 3,058 OCT B-scans captured from 163 participants. Model One achieved an area under the curve (AUC) of 0.912 [95% confidence interval (CI), 0.895–0.962], while Model Two achieved an overall AUC of 0.989 [95% CI, 0.982–0.993] for detecting pathological retinas from healthy samples. Moreover, Model Two achieved an average retinopathies detection accuracy of 94.52%. Heat maps showed that the algorithm focused on the area with pathological changes during processing, similar to manual grading in daily clinical work. Conclusions The proposed domain adaptation model showed a strong ability in reducing the domain distance between different OCT datasets.

Evaluating Changes in the Multiyear Predictability of the Pacific Decadal Oscillation Using Model Analogs since 1900

In this study, we investigate the changes in the multiyear predictability of the Pacific decadal oscillation (PDO) since 1900. A newly developed method, the model-analog method, is used to hindcast the PDO from 1900 to 2015. Model-analog hindcast provides comparable PDO prediction skills with the traditional assimilation-initialized forecast but with much lower computational costs. Our results show that PDO prediction skills have significantly changed over time. Specifically, the PDO could skillfully be predicted up to 5 years in advance in 1910–1960, but it can only be predicted 2–3 years in advance after 1960. We attribute these changes to the strength of the re-emergence process, where deep oceanic mixed layer temperature anomalies from one winter reappear in the following winters. In the high-prediction-skill period, the stronger re-emergence process provides more predictable information, leading to enhanced multiyear prediction skills. Our study offers new insights into the multiyear predictability of the PDO over a long-time frame and sheds light on the reasons behind the low prediction skill of the PDO.

Oceanic bottom mixed layer in the Clarion-Clipperton Zone: potential influence on deep-seabed mining plume dispersal

The oceanic bottom mixed layer (BML) is a well mixed, weakly stratified, turbulent boundary layer. Adjacent to the seabed, the BML is of intrinsic importance for studying ocean mixing, energy dissipation, particle cycling and sediment-water interactions. While deep-seabed mining of polymetallic nodules is anticipated to commence in the Clarion-Clipperton Zone (CCZ) of the northeastern tropical Pacific Ocean, knowledge gaps regarding the form of the BML and its potentially key influence on the dispersal of sediment plumes generated by deep-seabed mining activities are yet to be addressed. Here, we report recent field observations from the German mining licence area in the CCZ that characterise the structure and variability of the BML locally. Quasi-uniform profiles of potential temperature extending from the seafloor reveal the presence of a spatially and temporally variable BML with an average local thickness of approximately 250 m. Deep horizontal currents in the region have a mean speed of 3.5 cm s-1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{-1}$$\\end{document} and a maximum speed of 12 cm s-1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{-1}$$\\end{document} at 18.63 ms above bottom over an 11 month record. The near-bottom currents initially have a net southeastward flow, followed by westward and southward flows with the development of complex, anticyclonic flow patterns. Theoretical predictions and historical data show broad consistency with mean BML thickness but cannot explain the observed heterogeneity of local BML thickness. We postulate that deep pressure anomalies induced by passing surface mesoscale eddies and abyssal thermal fronts could affect BML thickness, in addition to local topographic effects. A simplified transport model is then used to study the influence of the BML on the interplay between turbulent diffusion and sediment settling in the transport of deep-seabed mining induced sediment plumes. Over a range of realistic parameter values, the effects of BML on plume evolution can vary significantly, highlighting that resolving the BML will be a crucial step for accurate numerical modelling of plume dispersal.

A review of the geophysical knowledge of the Borborema Province, NE-Brazil, and tectonic implications

This review summarizes the results of nearly 50 years of application of geophysical data to study tectonic problems in Borborema Province (BP), Northeastern Brazil. Much information has been obtained from the interpretation of magnetic, gamma-ray spectrometry, gravity, seismological, magnetotelluric, deep seismic and heat flow data. The resulting picture is that, in comparison with other Brazilian geological domains, BP has a thin crust, a lithosphere with small effective elastic thickness, the presence of predominantly positive free-air gravity anomalies, besides the most expressive geoid anomaly, in addition to relatively high heat flow and intraplate seismicity. These facts reinforce the hypothesis that the BP lithosphere suffered a strong impact in Gondwana fragmentation. Predominantly positive free-air gravity anomalies reveal that part of the province has not yet reached isostatic equilibrium. Gravity data established the boundaries of the BP crustal block with neighboring blocks; in particular, the crustal boundary with the São Francisco Craton is located in the innermost regions of the marginal fold belt, tens of kilometers at the north of the conventional limit. Magnetic data allowed to infer the continuity of basement structures beneath sedimentary basins, in particular that BP continues to the Transbrasiliano Lineament, underneath the sedimentary of the Parnaíba Basin. In addition, enhanced magnetic maps made it possible assessing the amount of deformation associated with the major shear zones and identified a continental-scale Cretaceous dike system crossing the BP. Gamma-ray spectrometry data allowed extensive geological mapping of metasedimentary sequences, granitic intrusions and basement complexes. Seismological data allowed to recognize that major shear zones in BP have lithospheric significance. Deep seismic data evidenced a well-defined contrast between the upper and lower crust in BP (Conrad discontinuity). Magnetotelluric data revealed several deep geoelectrical anomalies in BP, some of them interpreted as relics of subduction zones. Intraplate seismicity studies allowed inferring that the current stress state in the BP upper crust is predominantly transcurrent. The list of open problems in BP is also expressive. Notable examples are: a) it has not yet been possible to define which type of orogeny (accretionary or intracontinental?) was dominant in the Neoproterozoic; b) the cause of the expressive geoid anomaly is unknown; c) the tectonic significance of the Transbrasiliano Lineament (collisional suture? Just an intracontinental shear zone?) is also unknown. Regarding economic applications, the use of geophysical data to prospect new ore deposits is still relatively poor. We hope that this review can contribute to increase the familiarity and acceptance of geophysical results by the geological community.

Electron transfer rules of minerals under pressure informed by machine learning

Electron transfer is the most elementary process in nature, but the existing electron transfer rules are seldom applied to high-pressure situations, such as in the deep Earth. Here we show a deep learning model to obtain the electronegativity of 96 elements under arbitrary pressure, and a regressed unified formula to quantify its relationship with pressure and electronic configuration. The relative work function of minerals is further predicted by electronegativity, presenting a decreasing trend with pressure because of pressure-induced electron delocalization. Using the work function as the case study of electronegativity, it reveals that the driving force behind directional electron transfer results from the enlarged work function difference between compounds with pressure. This well explains the deep high-conductivity anomalies, and helps discover the redox reactivity between widespread Fe(II)-bearing minerals and water during ongoing subduction. Our results give an insight into the fundamental physicochemical properties of elements and their compounds under pressure.

Identifying performance anomalies in fluctuating cloud environments: A robust correlative-GNN-based explainable approach

Cloud computing provides scalable and elastic resources to customers as a low-cost, on-demand utility service. Multivariate time series anomaly detection is crucial to promise the overall performance of cloud computing systems. However, due to the complexity and high dynamics of cloud environments, anomaly detections caused by irregular fluctuations in data and the robustness of models are challenging. To address these issues, we propose a deep learning-based anomaly detection method for multivariate time series for real-world operational clouds: Correlative-GNN with Multi-Head Self-Attention and Auto-Regression Ensemble Method (CGNN-MHSA-AR). Our method utilizes two parallel graph neural networks (GNN) to learn the time and feature inter-dependencies to achieve fewer false positives. Our approach leverages a multi-head self-attention, GRU, and AR model to capture multiple-dimensional information, leading to better detection robustness. CGNN-MHSA-AR can also provide an abnormal explanation based on the prediction error of its constituent univariate series. We compare the detection performance of CGNN-MHSA-AR with seven baseline methods on seven public datasets. The evaluation shows that the proposed CGNN-MHSA-AR outperforms its competitors with an F1-Score of 0.871 on average and is 19.9% better than state-of-the-art baseline methods. In addition, CGNN-MHSA-AR also offers to correctly identify the root cause of detected anomalies with up to 74.1% accuracy.

Feature Interaction-Based Reinforcement Learning for Tabular Anomaly Detection

Deep learning-based anomaly detection (DAD) has been a hot topic of research in various domains. Despite being the most common data type, DAD for tabular data remains under-explored. Due to the scarcity of anomalies in real-world scenarios, deep semi-supervised learning methods have come to dominate, which build deep learning models and leverage a limited number of labeled anomalies and large-scale unlabeled data to improve their detection capabilities. However, existing works share two drawbacks. (1) Most of them simply treat the unlabeled samples as normal ones, ignoring the problem of label contamination, which is very common in real-world datasets. (2) Only very few works have designed models specifically for tabular data instead of migrating models from other domains to tabular data. Both of them will limit the model’s performance. In this work, we propose a feature interaction-based reinforcement learning for tabular anomaly detection, FIRTAD. FIRTAD incorporates a feature interaction module into a deep reinforcement learning framework; the former can model tabular data by learning a relationship among features, while the latter can effectively exploit available information and fully explore suspicious anomalies from the unlabeled samples. Extensive experiments on three datasets not only demonstrate its superiority over the state-of-art methods but also confirm its robustness to anomaly rarity, label contamination and unknown anomalies.

A Study on the Effectiveness of Deep Learning-Based Anomaly Detection Methods for Breast Ultrasonography.

In the medical field, it is delicate to anticipate good performance in using deep learning due to the lack of large-scale training data and class imbalance. In particular, ultrasound, which is a key breast cancer diagnosis method, is delicate to diagnose accurately as the quality and interpretation of images can vary depending on the operator's experience and proficiency. Therefore, computer-aided diagnosis technology can facilitate diagnosis by visualizing abnormal information such as tumors and masses in ultrasound images. In this study, we implemented deep learning-based anomaly detection methods for breast ultrasound images and validated their effectiveness in detecting abnormal regions. Herein, we specifically compared the sliced-Wasserstein autoencoder with two representative unsupervised learning models autoencoder and variational autoencoder. The anomalous region detection performance is estimated with the normal region labels. Our experimental results showed that the sliced-Wasserstein autoencoder model outperformed the anomaly detection performance of others. However, anomaly detection using the reconstruction-based approach may not be effective because of the occurrence of numerous false-positive values. In the following studies, reducing these false positives becomes an important challenge.

A Quantum-Classical Hybrid Solution for Deep Anomaly Detection.

Machine learning (ML) has achieved remarkable success in a wide range of applications. In recent ML research, deep anomaly detection (AD) has been a hot topic with the aim of discriminating among anomalous data with deep neural networks (DNNs). Notably, image AD is one of the most representative tasks in current deep AD research. ML's interaction with quantum computing is giving rise to a heated topic named quantum machine learning (QML), which enjoys great prospects according to recent academic research. This paper attempts to address the image AD problem in a deep manner with a novel QML solution. Specifically, we design a quantum-classical hybrid DNN (QHDNN) that aims to learn directly from normal raw images to train a normality model and then exclude images that do not conform to this model as anomalies during its inference. To enable the QHDNN to perform satisfactorily in deep image AD, we explore multiple quantum layer architectures and design a VQC-based QHDNN solution. Extensive experiments were conducted on commonly used benchmarks to test the proposed QML solution, whose results demonstrate the feasibility of addressing deep image AD with QML. Importantly, the experimental results show that our quantum-classical hybrid solution can even yield superior performance to that of its classical counterpart when they share the same number of learnable parameters.

Deep Clustering-Based Anomaly Detection and Health Monitoring for Satellite Telemetry

Satellite telemetry data plays an ever-important role in both the safety and the reliability of a satellite. These two factors are extremely significant in the field of space systems and space missions. Since it is challenging to repair space systems in orbit, health monitoring and early anomaly detection approaches are crucial for the success of space missions. A large number of efficient and accurate methods for health monitoring and anomaly detection have been proposed in aerospace systems but without showing enough concern for the patterns that can be mined from normal operational telemetry data. Concerning this, the present paper proposes DCLOP, an intelligent Deep Clustering-based Local Outlier Probabilities approach that aims at detecting anomalies alongside extracting realistic and reasonable patterns from the normal operational telemetry data. The proposed approach combines (i) a new deep clustering method that uses a dynamically weighted loss function with (ii) the adapted version of Local Outlier Probabilities based on the results of deep clustering. The DCLOP approach effectively monitors the health status of a spacecraft and detects the early warnings of its on-orbit failures. Therefore, this approach enhances the validity and accuracy of anomaly detection systems. The performance of the suggested approach is assessed using actual cube satellite telemetry data. The experimental findings prove that the suggested approach is competitive to the currently used techniques in terms of effectiveness, viability, and validity.

Deep Domain-Adversarial Anomaly Detection With One-Class Transfer Learning

Despite the big success of transfer learning techniques in anomaly detection, it is still challenging to achieve good transition of detection rules merely based on the preferred data in the anomaly detection with one-class classification, especially for the data with a large distribution difference. To address this challenge, a novel deep one-class transfer learning algorithm with domain-adversarial training is proposed in this paper. First, by integrating a hypersphere adaptation constraint into domain-adversarial neural network, a new hypersphere adversarial training mechanism is designed. Second, an alternative optimization method is derived to seek the optimal network parameters while pushing the hyperspheres built in the source domain and target domain to be as identical as possible. Through transferring one-class detection rule in the adaptive extraction of domain-invariant feature representation, the end-to-end anomaly detection with one-class classification is then enhanced. Furthermore, a theoretical analysis about the model reliability, as well as the strategy of avoiding invalid and negative transfer, is provided. Experiments are conducted on two typical anomaly detection problems, i.e., image recognition detection and online early fault detection of rolling bearings. The results demonstrate that the proposed algorithm outperforms the state-of-the-art methods in terms of detection accuracy and robustness.