Professor Suman Kalia, Department of Computer Science and Information Sciences, Saint Peter's University, Jersey City, NJ
Presently, generative AI has taken center stage in the news media, educational institutions, and the world at large. Machine learning has been a decades-old phenomenon, with little exposure to the average person until very recently. In the natural world, the oldest and best example of a “generative” model is the human being - one can close one’s eyes and imagine several plausible different endings to one’s favorite TV show. This paper focuses on the impact of generative and machine learning AI on the financial industry. Although generative AI is an amazing tool for a discriminant user, it also challenges us to think critically about the ethical implications and societal impact of these powerful technologies on the financial industry. It requires ethical considerations to guide decision-making, mitigate risks, and ensure that generative AI is developed and used to align with ethical principles, social values, and in the best interests of communities.
Generative AI, Machine Learning, Deep Learning, Financial Industry.
Jingbo Jia, Peng Wu, and Hussain Dawood, School of Information Science and Engineering, University of Jinan, Jinan, Department of Information Engineering Technology, National Skills University, Islamabad,Islamabad
To address the problem of insufficient failure data generated by disks and the imbalance between the number of normal and failure data. The existing Conditional Tabular Generative Adversarial Networks(CTGAN) deep learning methods have been proven to be effective in solving imbalance disk failure data. But CTGAN cannot learn the internal information of disk failure data very well. In this paper, a fault diagnosis method based on improved CTGAN, a classifier for specific category discrimination is added and a discriminator generate adversarial network based on residual network is proposed. We named it Residual Conditional Tabular Generative Adversarial Networks (RCTGAN). Firstly, to enhance the stability of system a residual network is utilized. RCTGAN uses a small amount of real failure data to synthesize fake fault data; Then, the synthesized data is mixed with the real data to balance the amount of normal and failure data; Finally, four classifier (multilayer perceptron, support vector machine, decision tree, random forest) models are trained using the balanced data set, and the performance of the models is evaluated using G-mean. The experimental results show that the data synthesized by the RCTGAN can further improve the fault diagnosis accuracy of the classifier.
Imbalanced dataset, CTGAN, Synthesis data, Disk failure data, Classification.
Shuangshuang Yuan, Peng Wu, and Yuehui Chen, School of Information Science and Engineering, University of Jinan, 250022, Jinan, China
Data class imbalance is a common problem in classification problems, where minority class samples are often more important and more costly to misclassify in a classification task. Therefore, it is very important to solve the data class imbalance classification problem. The SMART dataset exhibits an evident class imbalance, comprising a substantial quantity of healthy samples and a comparatively limited number of defective samples. This dataset serves as a reliable indicator of the disc’s health status. In this paper, we obtain the best balanced disk SMART dataset for a specific classification model by mixing and integrating the data synthesised by multivariate generative adversarial networks (GAN) to balance the disk SMART dataset at the data level; and combine it with genetic algorithms to obtain higher disk fault classification prediction accuracy on a specific classification model.
Disk failure prediction, GAN, Genetic algorithm, Data imbalance, SMART data.
Guangfu Gao, Peng Wu, and Hussain Dawood, School of lnformation Science and Engineering, University of Jinan, Jinan Department of Information Engineering Technology, National Skills University, Islamabad, Islamabad
Large scale data storage is susceptible to failure. As disks are damaged and replaced, traditional machine learning models, which rely on historical data to make predictions, struggle to accurately predict disk failures. This paper presents a novel method for predicting disk failures by leveraging multi-layer domain adaptive learning techniques. First, disk data with numerous faults is selected as the source domain, and disk data with fewer faults is selected as the target domain. A training of the feature extraction network is performed with the selected origin and destination domains. The contrast between the two domains facilitates the transfer of diagnostic knowledge from the domain of source and target. According to the experimental findings, it has been demonstrated that the suggested approach possesses the capability to produce a dependable predictive model and enhance the accuracy in forecasting disk failures, even when only a limited number of failure samples are available.
Disk failure prediction, Transfer learning, Domain adaptation, Distance metric.
Shuangshuang Yuan, Peng Wu, Yuehui Chen, and Qiang Li, School of Information Science and Engineering, University of Jinan, Jinan 250022, China
Class imbalance exists in many classification problems, and since the data is designed for accuracy, imbalance in data classes can lead to classification challenges with a few classes having higher misclassification costs. The Backblaze dataset, a widely used dataset related to hard discs, has a small amount of failure data and a large amount of health data, which exhibits a serious class imbalance. In this paper, we summarise and sort out the research in the field of unbalanced data from three aspects, namely, data-level methods, algorithmic-level methods and hybrid methods, and summarise and analyse the problems, algorithmic ideas, strengths and weaknesses of each type of methods, as well as discussing the challenges of unbalanced data classification and strategies to deal with them.
Classification, Disk Failure Prediction, Imbalanced Dataset, Data Processing.
Ihab Sekhi, Institute of Information Technology, Miskolc University, 3515 Miskolc
Cloud computing is widely considered a transformative force in the computing world and is poised to replace the traditional office setup as an industry standard. However, given the relative novelty of these services and challenges such as the impact of physical distance on Round-Trip Time (RTT), questions have arisen regarding system performance and associated billing structures. The primary objective of this study is to address these concerns. We aim to alleviate doubts by leveraging a fuzzy logic system to classify distances between regions that support computing services and compare them with the conventional web hosting format. To achieve this, we analyse the responses of one of these services, like Amazon Web Services, across different distance categories (near, medium, and far) between regions and strive to conclude overall system performance. Our tests reveal that significant data is consistently lost during customer transmission despite exhibiting superior round-trip times. We delve into this issue and present our findings, which may illuminate the observed anomalous behaviour.
Round Trip Time, Wireless Network, SLA, cloud computing.
Zied Guitouni1, Aya Zairi2 and Mounir Zrigui2, 1Department of Electronic and Micro-Electronic Laboratory, Faculty of Sciences of Monastir, Tunisia, 2Informatics Department, Faculty of Sciences of Monastir, 5000, Tunisia
Cryptographic key generation plays a vital role in securing data transmission and device authentication in Internet of Things (IoT) systems. However, traditional random key generation methods face challenges in terms of security, efficiency and scalability, especially for resource-constrained IoT devices. Neural networks have emerged as a promising technique for generating cryptographic keys due to their ability to learn complex patterns. However, their architecture has a significant impact on performance. This paper provides a comparative analysis of three commonly used neural network architectures for cryptographic key generation on IoT devices. We propose a novel neural network-based key generation algorithm and implement it using each architecture. The models are trained to generate cryptographic keys of different sizes from random input data. Performance is evaluated based on key metrics like accuracy, loss, key randomness, and model complexity. Experimental results show that the FFNN architecture achieves over 99% accuracy and passes all randomness tests, outperforming the alternatives. CNNs underperform due to focusing on spatial features irrelevant for keys. RNNs struggle with keys' complex long-range dependencies.
Neural network architectures, cryptographic key generation algorithm, feedforward neural networks, convolutional neural networks, recurrent neural networks, evaluation for security in IoT devices .