EM-AUC: A Novel Algorithm for Evaluating Anomaly Based Network Intrusion Detection Systems DOI Creative Commons

Kevin Z. Bai,

John M. Fossaceca

Sensors, Год журнала: 2024, Номер 25(1), С. 78 - 78

Опубликована: Дек. 26, 2024

Effective network intrusion detection using anomaly scores from unsupervised machine learning models depends on the performance of models. Although do not require labels during training and testing phases, assessment their metrics evaluation phase still requires comparing against labels. In real-world scenarios, absence in massive datasets makes it infeasible to calculate metrics. Therefore, is valuable develop an algorithm that calculates robust without this paper, we propose a novel algorithm, Expectation Maximization-Area Under Curve (EM-AUC), derive Area ROC (AUC-ROC) Precision-Recall (AUC-PR) by treating unavailable as missing data replacing them through posterior probabilities. This was applied two datasets, yielding results. To best our knowledge, first time AUC-ROC AUC-PR, derived labels, have been used evaluate systems. The EM-AUC enables model training, testing, proceed comprehensive offering cost-effective scalable solution for selecting most effective detection.

Язык: Английский

Combining the Viterbi Algorithm and Graph Neural Networks for Efficient MIMO Detection DOI Open Access
Thien An Nguyen, Xuan-Toan Dang, Oh‐Soon Shin

и другие.

Electronics, Год журнала: 2025, Номер 14(9), С. 1698 - 1698

Опубликована: Апрель 22, 2025

In the advancement of wireless communication, multiple-input, multiple-output (MIMO) detection has emerged as a promising technique to meet high throughput requirements 6G networks. Traditionally, MIMO relies on conventional algorithms, such zero forcing and minimum mean square error, mitigate interference enhance desired signal. Mathematically, these algorithms operate linear transformations or functions received signals. To further performance, researchers have explored use nonlinear by leveraging deep learning structures models. this paper, we propose novel model that integrates Viterbi algorithm with graph neural network (GNN) improve signal in systems. Our approach begins detecting using VA, whose output serves initial input for GNN model. Within framework, are represented nodes, while channel structure defines edges. Through an iterative message-passing mechanism, progressively refines signal, enhancing its accuracy better approximate originally transmitted Experimental results demonstrate proposed outperforms existing approaches, leading superior performance.

Язык: Английский

Процитировано

0
Robust Text-to-Cypher Using Combination of BERT, GraphSAGE, and Transformer (CoBGT) Model DOI Creative Commons

Quoc-Bao-Huy Tran,

Aagha Abdul Waheed,

Sun-Tae Chung

и другие.

Applied Sciences, Год журнала: 2024, Номер 14(17), С. 7881 - 7881

Опубликована: Сен. 4, 2024

Graph databases have become essential for managing and analyzing complex data relationships, with Neo4j emerging as a leading player in this domain. Neo4j, high-performance NoSQL graph database, excels efficiently handling connected data, offering powerful querying capabilities through its Cypher query language. However, due to Cypher’s complexities, making it more accessible nonexpert users requires translating natural language queries into Cypher. Thus, paper, we propose text-to-Cypher model effectively translate In our proposed model, combine several methods enable interact using the English Our approach includes three modules: key-value extraction, relation–properties prediction, generation. For extraction leverage BERT GraphSAGE extract features from Finally, use Transformer generate these features. Additionally, lack of datasets, introduced new dataset that contains questions information within paired corresponding ground truths. This aids future learning, validation, comparison on task. Through experiments evaluations, demonstrate achieves high accuracy efficiency when comparing some well-known seq2seq such T5 GPT2, an 87.1% exact match score dataset.

Язык: Английский

Процитировано

1
Exploiting Extrinsic Information for Serial MAP Detection by Utilizing Estimator in Holographic Data Storage Systems DOI Creative Commons
Thien An Nguyen, Jaejin Lee

Applied Sciences, Год журнала: 2024, Номер 15(1), С. 139 - 139

Опубликована: Дек. 27, 2024

In the big data era, are created in huge volume. This leads to development of storage devices. Many technologies proposed for next generation fields. However, among them, holographic (HDS) has attracted much attention and been introduced as promising candidate meet increasing demand capacity speed. For signal processing, HDS faces two major challenges: inter-page interference (IPI) two-dimensional (2D) interference. To access IPI problem, we can use balanced coding, which converts user into an intensity level with uniformly distributed values each page. 2D interference, equalizer detection mitigate often-used methods wireless communication only handle one-dimensional (1D) signal. Thus, combine equalizer, detection, estimator reduce 1D this paper, a combined model using serial maximum posteriori (MAP) improve systems. our model, instead Viterbi algorithm predict upper–lower (UPI) or left–right (LRI) converting received ISI, used extrinsic information MAP detection. preserves improves by information. The simulation results demonstrate that significantly bit-error rate (BER) performance compared previous studies.

Язык: Английский

Процитировано

1
Software Defined Network and Graph Neural Network-based Anomaly Detection Scheme for High Speed Networks DOI Creative Commons

Archan Dadhania,

Poojan Dave,

Jitendra Bhatia

и другие.

Cyber Security and Applications, Год журнала: 2024, Номер unknown, С. 100079 - 100079

Опубликована: Ноя. 1, 2024

Язык: Английский

Процитировано

0
EM-AUC: A Novel Algorithm for Evaluating Anomaly Based Network Intrusion Detection Systems DOI Creative Commons

Kevin Z. Bai,

John M. Fossaceca

Sensors, Год журнала: 2024, Номер 25(1), С. 78 - 78

Опубликована: Дек. 26, 2024

Effective network intrusion detection using anomaly scores from unsupervised machine learning models depends on the performance of models. Although do not require labels during training and testing phases, assessment their metrics evaluation phase still requires comparing against labels. In real-world scenarios, absence in massive datasets makes it infeasible to calculate metrics. Therefore, is valuable develop an algorithm that calculates robust without this paper, we propose a novel algorithm, Expectation Maximization-Area Under Curve (EM-AUC), derive Area ROC (AUC-ROC) Precision-Recall (AUC-PR) by treating unavailable as missing data replacing them through posterior probabilities. This was applied two datasets, yielding results. To best our knowledge, first time AUC-ROC AUC-PR, derived labels, have been used evaluate systems. The EM-AUC enables model training, testing, proceed comprehensive offering cost-effective scalable solution for selecting most effective detection.

Язык: Английский

Процитировано

0