Unusual Events Research Articles

Performance Measurement of a DMEWMA Control Chart on an AR(p) Model with Exponential White Noise

The double-modified exponentially weighted moving average (DMEWMA) control chart running an autoregressive (AR) process is proposed to detect unusual events. The AR equation and the DMEWMA statistic are combined to evaluate the control limit of the exponential residual term to obtain the explicit formula for the average run length (ARL). The ARLs computed using the explicit formula approach and the well-established numerical integral equation method were compared to validate the former. The efficiencies of the original EWMA, MEWMA, and DMEWMA control charts running AR processes based on simulation and real data were compared by using the results of ARL and relative mean index calculations. The results indicate that the explicit formula for the ARL of an AR process running on a double-modified EWMA control chart detected changes more quickly than on either of the other two control charts for small and moderate changes. Finally, real data on COVID-19 is provided to demonstrate the application of this explicit formula.

MORBIDITY AND MORTALITY PATTERNS OF INDIAN RIVER LAGOON COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS TRUNCATUS) 2002-2020.

Mortality patterns in cetaceans are critical to understanding population health. Common bottlenose dolphins (Tursiops truncatus truncatus) inhabiting the Indian River Lagoon (IRL), Florida have been subjected to four unusual mortality events (UMEs), highlighting the need to evaluate morbidity and mortality patterns. Complete gross examinations were conducted on 392 stranded dolphins and histopathological analyses were conducted for 178 animals (2002-2020). The probable causes of mortality were grouped by etiologic category: degenerative, metabolic, nutritional, inflammatory (infectious and noninfectious disease), and trauma. Probable cause of mortality was determined in 57% (223/392) of cases. Inflammatory disease (infectious/noninfectious) and trauma were the most common. Inflammatory disease accounted for 41% of cases (91/223), with the lungs (pneumonia) most commonly affected. Trauma accounted for 36% of strandings (80/223). The majority of trauma cases were due to anthropogenic activities (entanglement, fishing gear or other debris ingestion, and propeller strikes), accounting for 58% of trauma cases (46/80). Natural trauma (prey-associated esophageal obstruction or asphyxiation, shark bites, and stingray interactions) accounted for 12% of all cases (26/223), and trauma of undetermined origin was identified in 4% of cases (8/223). Starvation or inanition (nutritional) were the probable cause of mortality in 17% of cases and peaked during the 2013 UME (61% of cases). Degenerative and metabolic etiologies accounted for 5% of cases. This study represents the most comprehensive evaluation of morbidity and mortality patterns in IRL dolphins. Because IRL dolphins are routinely exposed to anthropogenic threats and have endured multiple UMEs, these baseline data are critical to the conservation and management of this population.

Domoic acid in Canadian Pacific waters, from 2016 to 2021, and relationships with physical and chemical conditions

Domoic acid, a phycotoxin produced by species of the marine diatom Pseudo-nitzschia, can cause deleterious impacts to marine food webs and human health. Domoic acid and Pseudo-nitzschia spp. were surveyed from 2016 to 2021 in the Pacific waters of Canada to assess their occurrences, concentrations, and relationships with physical and chemical conditions. Domoic acid was common, occurring in measurable concentrations in 73 % of the 454 samples. It occurred in all regions (west coast of Vancouver Island, Salish Sea, Queen Charlotte Sound / Hecate Strait, deep oceanic NE Pacific), in all years and all seasons. Median concentrations were highest along the west coast of Vancouver Island, and lowest in the oceanic waters of the NE Pacific. Winter had the lowest concentrations; no significant differences occurred between spring, summer, and autumn. High domoic acid concentrations equal to or above 100 ng/L were not common, occurring in about 5 % of samples, but in all seasons and all years except 2019. All six Pseudo-nitzschia taxa identified had similar median concentrations, but different frequencies of occurrence. P. cf. australis appeared to be the major contributor to high concentrations of domoic acid. Physico-chemical conditions were described by ten variables: temperature, salinity, density difference between 30 m and the surface (a proxy for vertical stability), chlorophyll a, nitrate, phosphate, silicate, and the ratios nitrate:phosphate, nitrate:silicate, and silicate:phosphate. Statistical analyses, using general linear models, of their relationships with the absence/presence of Pseudo-nitzschia spp. found silicate (negative) to be the most influential variable common in both the west coast of Vancouver Island and Salish Sea regions. Temperature and chlorophyll a were the most influential variables which determined the log10 abundance of Pseudo-nitzschia spp. in both regions. Analyses of the absence/presence of particulate domoic acid per Pseudo-nitzschia cell (excluding P. americana) found chlorophyll a to be the most influential variable common in both regions, whereas no common influential variable determined the log10 concentration of particulate domoic acid per Pseudo-nitzschia cell (excluding P. americana). These results were generally similar to those of other studies from this area, although this study extends these findings to all seasons and all regions of Canada's Pacific waters. The results provide important background information against which major outbreaks and unusual events can be compared. A domoic acid surveillance program during synoptic oceanographic surveys can help to understand where and when it reaches high concentrations at sea and the potential impacts to the marine ecosystem.

Proposing a new optimized forecasting model for the failure rate of power distribution network thermal equipment for educational centers

To gain a deep understanding of maintenance and repair planning, investigate the weak points of the distribution network, and discover unusual events, it is necessary to trace the shutdowns that occurred in the network. Many incidents happened due to the failure of thermal equipment in schools. On the other hand, the most important task of electricity distribution companies is to provide reliable and stable electricity, which minimal blackouts and standard voltage should accompany. This research uses seasonal time series and artificial neural network approaches to provide models to predict the failure rate of one of the equipment used in two areas covered by the greater Tehran electricity distribution company. These data were extracted weekly from April 2019 to March 2021 from ENOX incident registration software. For this purpose, after pre-processing the data, the appropriate final model was presented with the help of Minitab and MATLAB software. Also, average air temperature, rainfall, and wind speed were selected as input variables to the neural network. The mean square error has been used to evaluate the proposed models’ error rate. The results show that the time series models performed better than the multi-layer perceptron neural network in predicting the failure rate of the target equipment and can be used to predict future periods.

Anomaly detection in telemetry data using a jointly optimal one-class support vector machine with dictionary learning

Anomaly detection based on telemetry data is a major issue in satellite health monitoring, given that it can identify unusual or unexpected events to avoid serious accidents and ensure the safety and reliability of operations. This study proposes a jointly optimal one-class support vector machine with dictionary learning (JODL-OC) framework for multivariate anomaly detection in telemetry. Dictionary learning (DL) extracts the correlation and local dynamics of time series into a sparse matrix, and the one-class support vector machine (OCSVM) detects anomalies in the transformed sparse space. The optimisation objective of DL is modified by introducing a novel regularisation term called soft-boundary OCSVM, so that the JODL-OC seamlessly integrates DL, sparse representation, and classifier training into a unified model to jointly learn the features and decision boundary. Furthermore, we propose an iterative approach to solve the optimisation function and obtain the optimal JODL-OC framework for anomaly detection. A case analysis based on satellite antenna telemetry data reveals that the proposed method improves the detection precision, F1-score, FPR and FAR compared to other existing anomaly detection methods, particularly to illustrate the superior performance in correlation anomalies.

Using Distributed Temperature Sensing for Long-Term Monitoring of Pockmark Activity in the Gulf of Patras (Greece): Data Processing Hints and Preliminary Findings.

Patras Gulf pockmark field (Western Greece) is a tectonically controlled field that has been activated at least twice by strong earthquakes (M5.4, 14 July 1993 and M6.4, 8 June 2008), and episodic gas seepages have been recorded in the past using geophysical means. A distributed temperature sensor (DTS) system was deployed inside a shallow pockmark and along an active fault at the northern end of the field. This ongoing experiment represents the first long-term monitoring ever conducted on gas-bearing pockmarks and active faults by the DTS system. For now, we have acquired and analyzed data regarding about 1.56 years. One of the primary objectives of this study is to establish methodological queues for data processing and analysis, including spectral analysis and incomplete data treatment techniques, to be standardized for use in further stages of the experiment. Spectral analysis was proven capable of separating the temperature footprint of background environmental components, such as sea-atmosphere heat flux, tides, and winds/waves, from high-frequency temperature residuals. Those residuals represent unusual events that might be correlated to seismicity. Monitoring the causal relationship between seismic activity and seabed water temperature changes in the field was thus attempted. No significant local earthquakes occurred during the monitoring period. Although the relation between seismicity and irregular seabed water temperature events was not systematic, we postulate that four thermal events have a causative link with the local seismicity. The DTS system constitutes a low-cost monitoring system, and the promising preliminary results of this experiment suggest that it is worth testing for a longer period.

Anomaly Detection in a Smart Industrial Machinery Plant Using IoT and Machine Learning.

In an increasingly technology-driven world, the security of Internet-of-Things systems has become a top priority. This article presents a study on the implementation of security solutions in an innovative manufacturing plant using IoT and machine learning. The research was based on collecting historical data from telemetry sensors, IoT cameras, and control devices in a smart manufacturing plant. The data provided the basis for training machine learning models, which were used for real-time anomaly detection. After training the machine learning models, we achieved a 13% improvement in the anomaly detection rate and a 3% decrease in the false positive rate. These results significantly impacted plant efficiency and safety, with faster and more effective responses seen to unusual events. The results showed that there was a significant impact on the efficiency and safety of the smart manufacturing plant. Improved anomaly detection enabled faster and more effective responses to unusual events, decreasing critical incidents and improving overall security. Additionally, algorithm optimization and IoT infrastructure improved operational efficiency by reducing unscheduled downtime and increasing resource utilization. This study highlights the effectiveness of machine learning-based security solutions by comparing the results with those of previous research on IoT security and anomaly detection in industrial environments. The adaptability of these solutions makes them applicable in various industrial and commercial environments.

Unusually long path length for a nearly scatter-free solar particle event observed by Solar Orbiter at 0.43 au

Context. After their acceleration and release at the Sun, solar energetic particles (SEPs) are injected into the interplanetary medium and are bound to the interplanetary magnetic field (IMF) by the Lorentz force. The expansion of the IMF close to the Sun focuses the particle pitch-angle distribution, and scattering counteracts this focusing. Solar Orbiter observed an unusual solar particle event on 9 April 2022 when it was at 0.43 astronomical units (au) from the Sun. Aims. We show that the inferred IMF along which the SEPs traveled was about three times longer than the nominal length of the Parker spiral and provide an explanation for this apparently long path. Methods. We used velocity dispersion analysis (VDA) information to infer the spiral length along which the electrons and ions traveled and infer their solar release times and arrival direction. Results. The path length inferred from VDA is approximately three times longer than the nominal Parker spiral. Nevertheless, the pitch-angle distribution of the particles of this event is highly anisotropic, and the electrons and ions appear to be streaming along the same IMF structures. The angular width of the streaming population is estimated to be approximately 30 degrees. The highly anisotropic ion beam was observed for more than 12 h. This may be due to the low level of fluctuations in the IMF, which in turn is very probably due to this event being inside an interplanetary coronal mass ejection The slow and small rotation in the IMF suggests a flux-rope structure. Small flux dropouts are associated with very small changes in pitch angle, which may be explained by different flux tubes connecting to different locations in the flare region. Conclusions. The unusually long path length along which the electrons and ions have propagated virtually scatter-free together with the short-term flux dropouts offer excellent opportunities to study the transport of SEPs within interplanetary structures. The 9 April 2022 solar particle event offers an especially rich number of unique observations that can be used to limit SEP transport models.

The (im)possibility of complaint: on efforts of inverting and (en)countering the university

ABSTRACT Over the past decades, research has documented how endemic racism, sexism, and ableism are in academia. Universities have complaint procedures to address these issues. Much research focuses on individual experiences of making a complaint and the institutional uptake of complaints and demonstrates how such ‘isms’ are located in the individual rather than in the institution. This paper instead scrutinizes how complaint procedures mask and reproduce the structures with which complaints are concerned resulting in the complaints’ limited transformative abilities. I demonstrate how complaint procedures only allow for treating complaints as isolated, singular and unusual events that require temporary solutions, which ensures that complaints and complaint work are peripheralized while the white patriarchal ableist core of universities remains intact. Complaints as efforts of inverting the white patriarchal university are too limited as they are quickly reverted. Hence, what is needed is more than a mere procedure but a total inversion of the institution to make difference fit which requires work that goes in and beyond one’s institution.

Multivariate Time-Series Classification of Critical Events from Industrial Drying Hopper Operations: A Deep Learning Approach

In recent years, the advancement of Industry 4.0 and smart manufacturing has made a large amount of industrial process data attainable with the use of sensors installed on machines. This paper proposes an experimental predictive maintenance framework for an industrial drying hopper so that it can detect any unusual event in the hopper, which reduces the risk of erroneous fault diagnosis in the manufacturing shop floor. The experimental framework uses Deep Learning (DL) algorithms to classify Multivariate Time-Series (MTS) data into two categories—failure or unusual events and regular events—thus formulating the problem as a binary classification. The raw data extracted from the sensors contained missing values, suffered from imbalancedness, and were not labeled. Therefore, necessary preprocessing is performed to make them usable for DL algorithms and the dataset is self-labeled after defining the two categories precisely. To tackle the imbalanced data issue, data balancing techniques like ensemble learning with undersampling and Synthetic Minority Oversampling Technique (SMOTE) are used. Moreover, along with DL algorithms like Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM), Machine Learning (ML) algorithms like Support Vector Machine (SVM) and K-nearest neighbor (KNN) have also been used to perform a comparative analysis on the results obtained from these algorithms. The result shows that CNN is arguably the best algorithm for classifying this dataset into two categories and outperforms other traditional approaches as well as deep learning algorithms.

Propagation of tropical squall line-induced storm coastal inundation episodes in Java-Bali, Indonesia

The short-lived tropical squall lines could trigger weather-related hazards to the northern part of the Indonesia Maritime Continent (IMC), such as Sumatra and Kalimantan. Herein, we investigated the rare propagation event of the long-lived Sumatra squall line associated with a severe storm surge that induced coastal inundation in Java-Bali with devastating impacts from 22 May–2 June 2020. With a comprehensive approach combining observational, numerical, and analytical studies, for the first time, we proposed the possible mechanism related to the long-lived squall line over the IMC, which represents the largest equatorial tropical region with the most complicated air-sea interaction area in the world. Our findings suggest that the long-lived squall line related to the supercell-like thunderstorm initiated from multicell over central Sumatra on May 20, 2020, continuously propagated southeastward until several days later reached Bali. The near-quasi steady convective line has 6 hours of time travel from central Sumatra to west Java. The supercell-like rapidly develops from multicell with a deep convective updraft under the strong and fast cold pool (∼13.8 m s−1). The further southeastward propagation of squall line with broken line type seems reinforced by low-level moist transport from the Java Sea. This study also suggested that this unusual event of a long-lived squall line might occur more frequently in the warming upper ocean in the IMC.

682 Accidental Toothpick Ingestion Causing Liver Abscess – a Case Report and Review of Literature

Abstract Introduction Perforation due to foreign body ingestion leading to pyogenic liver abscess is an unusual but fatal event. We present a rare case of accidentally ingested toothpick that perforated the duodenum and migrated to the liver causing abscess. Case Presentation A 40-year-old gentleman presented with symptoms and signs of sepsis. He reported having eaten dry fish and chicken with bones 4 days prior to presentation. Abdominal computed tomography (CT) scan revealed hepatic collection containing a linear object in the middle. Non-operative management improved patient’s general condition but an increase in the size of collection lead to operative management. Discussion Usual course of formation of a pyogenic liver abscess is due to ascending cholangitis or infections causing hematogenous spread. Such an event due to a foreign body perforating the gastro-intestinal tract and migrating to liver is rarely seen. Presentation ranges from mild abdominal pain to full blown septicemia. Diagnosis is challenging as patients may present with vague symptoms, or sometimes don’t even remember having ingested an indigestible object. A trial of conservative management can be given to stable patients who improve, but deterioration necessitates surgical intervention. Conclusions Pyogenic liver abscess due to foreign body ingestion leading to intestinal perforation and displacing to liver is seldom encountered, and a toothpick leading to such an event is a very rare event. Patients may improve clinically with non-operative management; definitive management lies in the surgical removal of foreign body.

Differential diagnosis of sleep laughter: A case report and literature review

Sleep laughter is a relatively common phenomenon. It is classically seen during REM sleep, which is associated with dreams, and may be a component of REM sleep without atonia (RWA) as seen in cases of REM sleep behavior disorder (RBD). However, repetitive laughter episodes during NREM or during sleep-wake transition have not been described in the literature.We present a case of paroxysmal laughter out of drowsiness and NREM sleep, occurring almost every night, prompting evaluation for a possible seizure disorder. Multiple tests were unrevealing, including brain magnetic resonance imaging, polysomnogram, multiple sleep latency test and electroencephalogram. However, despite the lack of diagnostic certainty, this case provoked a discussion of key factors distinguishing parasomnia from seizure, which is useful for all physicians who may be faced with a case of unusual behavior in sleep.This case highlights the challenges that are encountered when trying to classify certain unusual sleep-related paroxysmal events.

Gut bacterial communities in Atlantic bottlenose dolphins (Tursiops truncatus) throughout a disease-driven (Morbillivirus) unusual mortality event.

Gut microbiomes are important determinants of animal health. In sentinel marine mammals where animal and ocean health are connected, microbiome impacts can scale to ecosystem-level importance. Mass mortality events affect cetacean populations worldwide, yet little is known about the contributory role of their gut bacterial communities to disease susceptibility and progression. Here, we characterized bacterial communities from fecal samples of common bottlenose dolphins, Tursiops truncatus, across an unusual mortality event (UME) caused by dolphin Morbillivirus (DMV). 16S rRNA gene sequence analysis revealed similar diversity and structure of bacterial communities in individuals stranding before, during, and after the 2013-2015 Mid-Atlantic Bottlenose Dolphin UME and these trends held in a subset of dolphins tested by PCR for DMV infection. Fine-scale shifts related to the UME were not common (10 of 968 bacterial taxa) though potential biomarkers for health monitoring were identified within the complex bacterial communities. Accordingly, acute DMV infection was not associated with a distinct gut bacterial community signature in T. truncatus. However, temporal stratification of DMV-positive dolphins did reveal changes in bacterial community composition between early and late outbreak periods, suggesting that gut community disruptions may be amplified by the indirect effects of accumulating health burdens associated with chronic morbidity.

The Stability and Growth Pact’s “Unusual Events” Clause and the Funding of the Refugee Crisis in the European Union

ABSTRACT The European Union (EU) has been tested by three great external events, the Syrian refugee crisis, the Coronavirus, and the Russian invasion of Ukraine. Each of these events challenges the EU’s Stability and Growth Pact (SGP) that sanctions member states incurring excessive budgetary deficits. This research focuses on the application of these rules to the refugee crisis that generated a dramatic and unpredicted surge in migrants and refugees who crossed into and applied for asylum in Europe. What was critical for the member states was how they could respond to the SGP to fund the cost of aiding and housing these immigrants. These rules were later applied in 2020 and 2022 to the budgetary costs incurred by the Coronavirus and the invasion of Ukraine.

Forecasting of an unusual dust event over western India by the Air Quality Early Warning System

This study presents the successful implementation of an operational Air Quality Early Warning System (AQEWS) in forecasting an unprecedented dust storm event that occurred in the western part of India from 21–24 January 2022. The AQEWS generates a daily 72-h forecast using the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) at the core, which is initialized with assimilated data from the Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth (AOD) retrievals and in-situ measurements of fine particulate matter (PM2.5). The AQEWS was able to forecast the outbreak of this dust storm reasonably well over the north of the Arabian Sea on 21 January, as well as its movement and dispersal towards the Indian landmass in the subsequent 72 h, resulting in a significant amount of transported dust that engulfed the west Indian region on 23–24 January. We utilize satellite observations and in situ measurements to quantify the ability of the forecasting system to predict the spatiotemporal characteristics of the dust storm. Notably, the AQEWS aerosol data assimilation process played a critical role in forecasting the anomalous increase of PM10 in several west Indian cities, as it significantly improved the initial conditions of PM10 by approximately 24–40 μg/m3 and PM2.5 concentration by approximately 30–38 μg/m3. Without assimilation, the system fails to capture the intensity, and pattern of variability of the dust storm. The mean bias of the 72-h prediction of PM10 was substantially reduced with assimilation, from −44% to 0.35% in Mumbai and from −57% to 13% in Gujarat. In selected regions, in situ measurements confirmed that the AQEWS adequately captured the peak intensity and pattern of the dust storm event. The AQEWS thus demonstrated improved forecasting skills in predicting an extreme dust event during winter and the adoption of chemical data assimilation enhanced the accuracy of the real-time air quality forecasting system.

A radiative transfer deep learning model coupled into WRF with a generic fortran torch adaptor

Advances in deep learning have created new opportunities for improving traditional numerical models. As the radiation parameterization scheme is crucial and time-consuming in numerical models, researchers sought to replace it with deep learning emulators. However, progress has been hindered at the offline emulation stage due to the technical complexity of the implementation. Additionally, the performance of the emulators when coupled with large-scale numerical models has yet to be verified. In this paper, we have developed a new tool called the Fortran Torch Adaptor (FTA) to facilitate this process and coupled deep learning emulators into the WRF model with it. The performance of various structured AI models was tested in terms of accuracy, generalization ability, and efficiency in different weather forecasting scenarios. Our findings revealed that deep learning models outperformed ordinary feedforward neural networks (FNN), achieving greater accuracy both online and offline, and leading to better overall forecasting results. When it came to unusual extreme weather events, all models were affected to some extent, but deep learning models exhibited less susceptibility than other models. With the assistance of FTA, deep learning models on GPU could achieve significant acceleration, ranging from 50x to 300x depending on the parameterization scheme replacing strategy. In conclusion, this research is crucial for both the theoretical and practical development of radiation transfer deep learning emulators. It demonstrates the emerging potential for using deep learning-based parameterizations in operational forecasting models.

Analysis of events reported from industrial radiation facilities

Radiation sources have multifarious societal applications in medical, industrial and research. Prevention of any untoward event is one of the important safety principles to be considered during the handling of radiation sources. However, some unusual events may occur during the handling of radiation sources, which may be attributed to various factors such as equipment malfunction, human error etc. Atomic Energy Regulatory Board (AERB) has a well-established mechanism for event reporting by the stakeholders. An analysis of the events reported by the facilities handling radiation sources for industrial radiography, nucleonic gauges and well-logging applications has been carried out. Analysis shows that the major contributing factors for reported events in industrial radiography are human error and equipment failure, whereas for nucleonic gauges and well logging facilities dominating factors are security-related issues and external (environmental) factors respectively. These reported events have not led to any serious accident, however, reported by the facilities as deviations from standard operating procedures are observed.

Epidemiological and genetic analysis of Cetacean Morbillivirus circulating on the Italian coast between 2018 and 2021.

Cetacean morbillivirus (CeMV) has caused several outbreaks, unusual mortality events, and interepidemic single-lethal disease episodes in the Mediterranean Sea. Since 2012, a new strain with a northeast (NE) Atlantic origin has been circulating among Mediterranean cetaceans, causing numerous deaths. The objective of this study was to determine the prevalence of CeMV in cetaceans stranded in Italy between 2018 and 2021 and characterize the strain of CeMV circulating. Out of the 354 stranded cetaceans along the Italian coastlines, 113 were CeMV-positive. This prevalence (31.9%) is one of the highest reported without an associated outbreak. All marine sectors along the Italian coastlines, except for the northern Adriatic coast, reported a positive molecular diagnosis of CeMV. In one-third of the CeMV-positive cetaceans submitted to a histological evaluation, a chronic form of the infection (detectable viral antigen, the absence of associated lesions, and concomitant coinfections) was suspected. Tissues from 24 animals were used to characterize the strain, obtaining 57 sequences from phosphoprotein, nucleocapsid, and fusion protein genes, which were submitted to GenBank. Our sequences showed the highest identity with NE-Atlantic strain sequences, and in the phylogenetic study, they clustered together with them. Regarding age and species, most of these individuals were adults (17/24, 70.83%) and striped dolphins (19/24, 79.16%). This study improves our understanding on the NE-Atlantic CeMV strain in the Italian waters, supporting the hypothesis of an endemic circulation of the virus in this area; however, additional studies are necessary to deeply comprehend the epidemiology of this strain in the Mediterranean Sea.

AD-Graph: Weakly Supervised Anomaly Detection Graph Neural Network

The main challenge faced by video-based real-world anomaly detection systems is the accurate learning of unusual events that are irregular, complicated, diverse, and heterogeneous in nature. Several techniques utilizing deep learning have been created to detect anomalies, yet their effectiveness on real-world data is often limited due to the insufficient incorporation of motion patterns. To address these problems and enhance the traditional functionality of anomaly detection systems for surveillance video data, we propose a weakly supervised graph neural-network-assisted video anomaly detection framework called AD-Graph. To identify temporal information from a series of frames, we extract 3D visual and motion features and represent these in a language-based knowledge graph format. Next, a robust clustering strategy is applied to group together meaningful neighbourhoods of the graph with similar vertices. Furthermore, spectral filters are applied to these graphs, and spectral graph theory is used to generate graph signals and detect anomalous events. Extensive experimental results over two challenging datasets, UCF-Crime and ShanghaiTech, show improvements of 0.35% and 0.78% against a state-of-the-art model.