Generalized support vector data description for anomaly detection

Turkoz, Mehmet; Kim, Sangahn; Son, Youngdoo; Jeong, Myong K.; Elsayed, Elsayed A.

doi:10.1016/j.patcog.2019.107119

Cited by 48 publications

(16 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A related challenge is formed by those species that are completely absent from the reference database on which the deep learning models are trained. Detecting such species requires techniques developed for multiple-class novelty/anomaly detection or open set/world recognition (94,95). A recent survey introduced various open set recognition methods with the two main approaches being discriminative and generative (96).…”

Section: Potential Deep Learning Applications In Entomologymentioning

confidence: 99%

Deep learning and computer vision will transform entomology

Høye

Ärje

Bjerge

et al. 2021

Proc. Natl. Acad. Sci. U.S.A.

305

206

View full text Add to dashboard Cite

Most animal species on Earth are insects, and recent reports suggest that their abundance is in drastic decline. Although these reports come from a wide range of insect taxa and regions, the evidence to assess the extent of the phenomenon is sparse. Insect populations are challenging to study, and most monitoring methods are labor intensive and inefficient. Advances in computer vision and deep learning provide potential new solutions to this global challenge. Cameras and other sensors can effectively, continuously, and noninvasively perform entomological observations throughout diurnal and seasonal cycles. The physical appearance of specimens can also be captured by automated imaging in the laboratory. When trained on these data, deep learning models can provide estimates of insect abundance, biomass, and diversity. Further, deep learning models can quantify variation in phenotypic traits, behavior, and interactions. Here, we connect recent developments in deep learning and computer vision to the urgent demand for more cost-efficient monitoring of insects and other invertebrates. We present examples of sensor-based monitoring of insects. We show how deep learning tools can be applied to exceptionally large datasets to derive ecological information and discuss the challenges that lie ahead for the implementation of such solutions in entomology. We identify four focal areas, which will facilitate this transformation: 1) validation of image-based taxonomic identification; 2) generation of sufficient training data; 3) development of public, curated reference databases; and 4) solutions to integrate deep learning and molecular tools.

show abstract

Section: Potential Deep Learning Applications In Entomologymentioning

confidence: 99%

Deep learning and computer vision will transform entomology

Høye

Ärje

Bjerge

et al. 2021

Proc. Natl. Acad. Sci. U.S.A.

305

206

View full text Add to dashboard Cite

show abstract

“…To compare the performance of different procedures, we use two different performance measures which are introduced by Turkoz et al (2020). The total classification accuracy rate (TCAR) is used for determination of the best parameters with validation dataset and the classification accuracy rate for the testing dataset which is defined as follows:…”

Section: Experimental Methodologymentioning

confidence: 99%

Multi-class Bayesian support vector data description with anomalies

Turkoz

Kim

2021

Ann Oper Res

Self Cite

View full text Add to dashboard Cite

Support vector data description (SVDD) procedure fits a spherically shaped boundary around the normal data by minimizing the volume of the description. However, the SVDD may not find an efficient boundary if the normal data consist of multiple classes. In addition to the multi-class normal data, some anomaly observations can be available. We propose a generalized SVDD procedure which finds multiple spheres around the multi-class data by incorporating the anomaly observations into the training procedure. Thus, descriptions for each class include as many as their corresponding class observations by keeping the other class and anomaly observations as far as possible. Moreover, we introduce a generalized Bayesian framework which utilizes the relationships among the classes by not only considering the prior information from normal classes but also the anomaly class. Experiments with various simulation studies and real-life applications demonstrate that the proposed approach can effectively identify the anomalies in multi-class data.

show abstract

“…Several ADS [7], [8] are used in recent IoT systems to detect anomalies and malicious nodes in the network and provide countermeasures to mitigate them. Consequently, they are considered as a solution to improve the system reliability.…”

Section: B Cloud and Fog Computingmentioning

confidence: 99%

READ-IoT: Reliable Event and Anomaly Detection Framework for the Internet of Things

Yahyaoui

Abdellatif²,

Yangui³

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Internet of Things (IoT) enables a myriad of applications by interconnecting software to physical objects. The objects range from wireless sensors to robots and include surveillance cameras. The applications are often critical (e.g. physical intrusion detection, fire fighting) and latency-sensitive. On the one hand, such applications rely on specific protocols (e.g. MQTT, COAP) and the network to communicate with the objects under very tight timeframe. On the other hand, anomalies (e.g. communication noise, sensors' failures, security attacks) are likely to occur in open IoT systems and can result by sending false alerts or the failure to properly detect critical events. To address that, IoT systems have to be equipped with anomaly detection processing in addition to the required event detection capability. This is a key feature that enables reliability and efficiency in IoT. However, anomaly detection systems can be themselves object of failures and attacks, and then can easily fall short to accomplish their mission. This paper introduces a Reliable Event and Anomaly Detection Framework for the Internet of Things (READ-IoT for short). The designed framework integrates events and anomalies detection into a single and common system that centralizes the management of both concepts. To enforce its reliability, the system relies on a reputationaware provisioning of detection capabilities that takes into account the vulnerability of the deployment hosts. As for validation, READ-IoT was implemented and evaluated using two real life applications, i.e. a fire detection and an unauthorized person detection applications. Several scenarios of anomalies and events were conducted using NSL-KDD public dataset, as well as, generated data to simulate routing attacks. The obtained results and performance measurements show the efficiency of READ-IoT in terms of event detection accuracy and real-time processing.

show abstract

Generalized support vector data description for anomaly detection

Cited by 48 publications

References 34 publications

Deep learning and computer vision will transform entomology

Deep learning and computer vision will transform entomology

Multi-class Bayesian support vector data description with anomalies

READ-IoT: Reliable Event and Anomaly Detection Framework for the Internet of Things

Contact Info

Product

Resources

About