Adaptive Machine Learning Algorithms for Real-Time Detection of Zero-Day Vulnerabilities

The rapid adoption of machine learning (ML) has increased the computational and storage demands on high-performance computing (HPC) and cloud datacenters. Unfortunately, deploying ML models on HPC and cloud platforms is challenging because it is often difficult to manage, financially costly, and poses substantial environmental concerns. This dissertation attempts to answer a simple but ambitious question: can we deploy ML models to meet the Quality of Service (QoS) requirements while remaining cost-effective and environmentally sustainable? The dissertation demonstrates that achieving this balance is indeed possible, albeit through employing non-intuitive strategies and user engagements. We show that a combination of fast and slow hardware can outperform the common strategy of exclusively employing the latest and fastest hardware when co-optimizing for both performance and cost. Realizing this potential, however, requires a strategy to create an effective hardware mix, a process rigorously developed within this dissertation. This dissertation further reveals that the carbon footprint of ML models can be managed by opportunistically trading off answer quality transiently and re-configuring/partitioning the GPU hardware resources, depending on the carbon intensity of available energy sources in the datacenter's region. Finally, this dissertation discovers that ML deployments can be made more environmentally sustainable by strategically engaging users and tuning response brevity, especially in the context of large language models (LLMs). The scientific advancements made in this dissertation aim to relieve the community from the cost constraints, hardware inefficiencies, and potential negative environmental impacts of large-scale deployment of ML models on HPC and cloud platforms. We are hopeful that its open-source contributions will promote further innovation and broader engagement toward sustainable ML deployment.--Author's abstract

Software fault detection is crucial for ensuring reliable and high-quality software systems. However, traditional fault detection methods often rely on manual inspection or rule-based techniques, which are time-consuming and prone to human errors. In this research, the researchers propose an enhanced fault detection approach using an adaptive neural transfer learning algorithm. The goal is to leverage the power of neural networks and adaptability to improve fault detection accuracy and classification performance. The problem addressed in this research is the need for more effective fault detection methods that can handle the complexities of modern software systems. Existing fault detection techniques lack adaptability and struggle to cope with diverse software scenarios. Neural networks have shown promise in pattern recognition and classification tasks, making them suitable for fault detection. However, fixed architectures and training strategy limit their performance in different software contexts. To address this problem, the research proposes an adaptive neural transfer learning algorithm for fault detection. The algorithm dynamically adjusts its neural network architecture and training process based on the characteristics of the software under test. It incorporates adaptive mechanisms, such as adjusting learning rates and regularization techniques, to optimize performance. Real-time feedback and performance evaluation during the training process drive the adaptive mechanisms. To evaluate the proposed approach, the researchers conducted a series of experiments using diverse software systems and fault scenarios. The research compared the performance of the adaptive algorithm with traditional fault detection methods, including rule-based techniques and fixed neural network architectures. Evaluation metrics such as accuracy, precision, recall, and F1 score were used. The results consistently show that the adaptive neural transfer learning algorithm outperforms existing methods, achieving higher fault detection accuracy and improved classification performance.

Artificial Intelligence (AI) and Machine Learning (ML) have experienced rapid growth in both industry and academia. However, the current ML and AI models demand significant computing and processing power to achieve desired accuracy and results, often restricting their use to high-capability devices. With advancements in embedded system technology and the substantial development in the Internet of Things (IoT) industry, there is a growing desire to integrate ML techniques into resource-constrained embedded systems for ubiquitous intelligence. This aspiration has led to the emergence of TinyML, a specialized approach that enables the deployment of ML models on resource-constrained, power-efficient, and low-cost devices. Despite its potential, the implementation of ML on such devices presents challenges, including optimization, processing capacity, reliability, and maintenance. This article delves into the TinyML model, exploring its background, the tools that support it, and its applications in advanced technologies. By understanding these aspects, we can better appreciate how TinyML is transforming the landscape of AI and ML in embedded and IoT systems.

The application of machine learning (ML) techniques for understanding and predicting organic matter (OM) and harmful algal blooms (HABs) in freshwater systems has increased significantly with the availability of abundant data and advanced monitoring technologies. However, there is a lack of comprehensive reviews concentrating on practical applications and delving into the potential risks associated with misrepresentation or inflation in constructing ML models. This review aims to bridge these gaps by providing a comprehensive overview of various aspects of ML applications in the context of OM and HABs in freshwater systems. It covers practical ML applications for rapid assessment, early warning, and driver analysis, highlighting the diverse range of techniques employed in these areas. Furthermore, it discusses the challenges and considerations associated with data handling, including using in situ and remote sensing data and the importance of appropriate data-splitting techniques to avoid data leakage. To ensure unbiased and reproducible results, this review offers recommendations for model improvement, such as utilizing explainable ML techniques to gain insights into model behavior and avoiding overreliance on a single ML algorithm. It also emphasizes the significance of deploying ML models through user-friendly interfaces, enabling non-experts in ML to effectively utilize these models in real-world water environments.

Machine learning (ML) is emerging as a revolutionary force in the evolution of self-service technologies (SSTs), transforming their design, implementation, and delivery across various industries. These self-organizing systems leverage ML to autonomously learn from interactions, process vast volumes of data, and make intelligent decisions without explicit programming. This chapter explores the pivotal role of artificial intelligence (AI) in advancing SSTs, focusing on key applications such as customer service chatbots, automated kiosks, and intelligent virtual assistants. In conclusion, this chapter emphasizes how ML is revolutionizing the SSTs' future development. The use of NLP, predictive analytics and adaptive learning is making the existing ML solutions accurate, efficient and highly personalized as never before. Further improving data credibility, privacy, and ethics issues will enhance the development of these systems to meet our needs and aspirations and continue to strengthen them as the foundations of global business and progress.

Background: Maternal depression and anxiety are significant public health concerns that play an important role in the health and well-being of mothers and children. The COVID-19 pandemic, the consequential lockdowns and related safety restrictions worldwide negatively affected the mental health of pregnant and postpartum women. Methods: This regional study aimed to develop a machine learning (ML) model for the prediction of maternal depression and anxiety. The study used a dataset collected from five Arab countries during the COVID-19 pandemic between July to December 2020. The population sample included 3569 women (1939 pregnant and 1630 postpartum) from five countries (Jordan, Palestine, Lebanon, Saudi Arabia, and Bahrain). The performance of seven machine learning algorithms was assessed for the prediction of depression and anxiety symptoms. Results: The Gradient Boosting (GB) and Random Forest (RF) models outperformed other studied ML algorithms with accuracy values of 83.3% and 83.2% for depression, respectively, and values of 82.9% and 81.3% for anxiety, respectively. The Mathew's Correlation Coefficient was evaluated for the ML models; the Naïve Bayes (NB) and GB models presented the highest performance measures (0.63 and 0.59) for depression and (0.74 and 0.73) for anxiety, respectively. The features' importance ranking was evaluated, the results showed that stress during pregnancy, family support, financial issues, income, and social support were the most significant values in predicting anxiety and depression. Conclusion: Overall, the study evidenced the power of ML models in predicting maternal depression and anxiety and proved to be an efficient tool for identifying and predicting the associated risk factors that influence maternal mental health. The deployment of machine learning models for screening and early detection of depression and anxiety among pregnant and postpartum women might facilitate the development of health prevention and intervention programs that will enhance maternal and child health in low- and middle-income countries.

BackgroundAugmented reality, computer vision and other (e.g. network functions, Internet-of-Things (IoT)) use cases can be realised in edge computing environments with machine learning (ML) techniques. For realisation of the use cases, it has to be understood how data is collected, stored, processed, analysed, and visualised in big data systems. In order to provide services with low latency for end users, often utilisation of ML techniques has to be optimized. Also, software/service developers have to understand, how to develop and deploy ML models in edge computing environments. Therefore, architecture design of big data systems to edge computing environments may be challenging.FindingsThe contribution of this paper is reference architecture (RA) design of a big data system utilising ML techniques in edge computing environments. An earlier version of the RA has been extended based on 16 realised implementation architectures, which have been developed to edge/distributed computing environments. Also, deployment of architectural elements in different environments is described. Finally, a system view is provided of the software engineering aspects of ML model development and deployment.ConclusionsThe presented RA may facilitate concrete architecture design of use cases in edge computing environments. The value of RAs is reduction of development and maintenance costs of systems, reduction of risks, and facilitation of communication between different stakeholders.

Interpretability in machine learning (ML) has become a critical focus, especially in the context of regulatory compliance and model risk management. As ML models are increasingly employed in high-stakes industries such as finance, healthcare, and insurance, understanding the decisions made by these models is paramount to ensuring compliance with regulatory requirements and managing associated risks. This paper explores the role of interpretability in improving the transparency, accountability, and fairness of machine learning models, highlighting its importance in meeting the standards set by regulatory bodies. It examines the various techniques used to enhance interpretability, including feature importance, surrogate models, and explainable AI (XAI) methods, and discusses their effectiveness in providing insights into model behavior. Furthermore, it emphasizes how interpretability aids in model risk management by enabling organizations to identify and mitigate potential biases, errors, and compliance violations. The paper also considers the challenges associated with interpretability, such as the trade-off between model complexity and transparency, and the need for consistent frameworks that align with evolving regulatory landscapes. In conclusion, the integration of interpretability within ML workflows not only fosters regulatory compliance but also enhances stakeholder trust and supports better decision-making in high-risk environments, thus playing a pivotal role in the responsible deployment of machine learning models.

With the advent of machine learning (ML) in the field of Materials Science, it has become obvious that trained models are limited by the amount and quality of the data used for training. Where researchers do not have access to the breadth and depth of labeled data that fields like image processing and natural language processing enjoy. In the specific application of materials discovery, there is the issue of continuity in atomistic datasets. Often if one relies on experimental data mined from literature and patents this data is only available for the most favorable of atomistic data. This ultimately leads to bias in the training dataset. In providing a solution, this research focuses on investigating the deployment of ML models trained on synthetic data and the development of a language-based approach for synthetically generating training datasets. It has been applied to three material science-related problems to prove these approaches work. The first problem was the prediction of dielectric properties, the second problem was the synthetic generation of chemical reaction datasets, and the third problem was the synthetic generation of quantum material datasets. All three applications proved successful and demonstrated the ability to generate continuous datasets that resolve the issue of dataset bias. This first study investigated the synthetic generation of complex dielectric properties of granular powders and their ability to train a ML network. The neural network was trained using a supervised learning approach and a common backpropagation. The network was double-validated using experimental data collected from a coaxial airline experiment. The second study demonstrated the synthetic generation of a chemical reaction database. An artificial intelligence model based on a Variational Autoencoder (VAE) has been developed and investigated to synthetically generate continuous datasets. The approach involves sampling the latent space to generate new chemical reactions that were assembled into the synthetic dataset. This developed technique is demonstrated by generating over 7,000,000 new reactions from a training dataset containing only 7,000 reactions. The generated reactions include molecular species that are larger and more diverse than the training set. The third study investigated a similar variational autoencoder approach to the second study but with the application of generating a synthetic dataset for quantum materials focusing on quantum sensing applications. The specific quantum sensors of interest are two-level quantum molecules that exhibit dipole blockade. This study offers an improved sampling algorithm by continuously feeding newly generated materials into a sampling algorithm to help

The review explores the application of machine learning (ML) to improve the accuracy of instrumentation in the oil and gas industry. The paper discusses the challenges faced in instrumentation accuracy and how ML can be utilized to address these challenges. It highlights the benefits of using ML, such as improved data accuracy, reduced maintenance costs, and enhanced operational efficiency. The review also covers the future prospects of ML in the oil and gas industry and concludes with a call to action for companies to adopt ML technologies to improve instrumentation accuracy. In the oil and gas industry, accurate instrumentation is crucial for ensuring safe and efficient operations. However, maintaining high levels of accuracy can be challenging due to factors such as environmental conditions, equipment aging, and human error. Machine learning (ML) offers a promising solution to enhance instrumentation accuracy by leveraging data-driven insights to improve monitoring and control systems. ML algorithms can analyze large volumes of data from various sensors and equipment to identify patterns and anomalies that may indicate potential issues. By continuously learning from new data, ML models can adapt to changing conditions and improve their accuracy over time. This proactive approach can help prevent equipment failures, minimize downtime, and optimize production processes. Furthermore, ML can also help reduce maintenance costs by enabling predictive maintenance strategies. By analyzing equipment performance data, ML models can predict when maintenance is likely to be needed, allowing operators to schedule maintenance activities proactively. This can help avoid costly unplanned downtime and reduce the need for unnecessary maintenance checks. Overall, leveraging ML to enhance instrumentation accuracy in oil and gas extraction offers significant benefits. It can improve operational efficiency, reduce costs, and enhance safety. As ML technologies continue to advance, the future prospects for enhancing instrumentation accuracy in the oil and gas industry look promising. Companies that embrace ML technologies stand to gain a competitive edge in the industry by improving their operational performance and reducing risks. Keywords: Leveraging, ML, Enhance, Instrumentation Accuracy, Oil and Gas Extraction.

Machine learning (ML) transitioned from a purely academic discipline to an applied field, gaining strategic importance in various industries. Meanwhile, Machine Learning Operations (MLOps) has been widely adopted by enterprises as a comprehensive approach for developing and managing machine learning applications. Despite its advantages, challenges remain. The rising demand for flexibility and scalability has led organizations to embrace multi-cloud and hybrid cloud architectures as preferred solutions. However, the autonomous and distributed nature of modern application development, combined with the complexity of training and deploying machine learning models, makes unified operational management impractical, and this will further affect application quality and efficiency. To address these challenges, this paper proposes a framework to manage model training and deployment in a multi-cloud environment. This framework uses a policy-based resource provisioning approach, agent-based application topology reconstruction, and a visualization dashboard. It aims to provide a cloud provider-neutral solution that enhances the quality of application operations. The framework design is introduced, followed by the implementation of a proof-of-concept prototype. Experiments conducted in various empirical scenarios demonstrate that the proposed framework effectively manages deployment resources while providing clear visibility and control across multiple clouds. The results confirm that this framework enhances control over deployment resources and optimizes model deployment efficiency in multi-cloud infrastructure.

Big data are data parallelism, not model parallelism. The sector that is investing the most in big data analytics in banking. Data have become an asset: They are key to the success of digital organizations. Machine learning is to computation what probability and statistics are to mathematics. Machine learning does not do magic: Without good data, it will not work and, the more data, the better; but data are not everything. Cloud computing allows combining measurement data from multiple remote devices, and the deployment of machine learning models without modifying the device or software at the edge. Work hard, work smart: Distinguish the important from the urgent. Do it with passion or not at all. Success is 99% failure. Equilibrium: Perseverance ↔ Flexibility. Once you stop learning, you start dying. Desire to be different. Be brave even if you are not: You fail until you get it. (Be) live in (you)self. The importance of networking. You do not get extra credit for doing it alone. Never underestimate the power of a mentor. The hard part is managing diversity. Society does not forgive errors but they are the engine that moves us forward.

Computing nodes in Edge Computing environments share unlimited data. Such data are exploited to locally build Machine Learning (ML) models for applications such as predictive analytics, exploratory analysis, and smart applications. This edge node-centric local learning reduces the need for data transfer and centralization, which is affected by different factors such as data privacy, data size, communication overhead, and computing resource limitations. Therefore, a collaborative learning fashion at the network edge has appeared as a promising paradigm that enables multiple distributed (edge) nodes to train and deploy ML models cooperatively without infringement of data privacy. Nevertheless, the variety, distribution and quality of data vary between edge nodes. Hence, selecting unsuitable edge nodes can have a negative impact on the ML model performances. We have devised (i) an intelligent node selection mechanism per analytics query based on the range of the availability of required training data at the edge and (ii) variants of collaborative learning processes engaging the most suitable nodes for models training and inference. We evaluate the efficiency of our selection mechanism and collaborative learning and provide a comparative assessment with other methods found in the literature using real data. The results showcase that our mechanism significantly outperforms baseline approaches and existing node selection mechanisms in distributed computing environments.

A suitable vulnerability assessment process improves the overall security of an organization, and therefore, reduces the chances that an attacker could breach the IT system. The detection of vulnerabilities is one of the first steps in software security lifecycle, but in some cases, it is not enough. Understanding how a vulnerability can affect the rest of the system as well as forecasting its exploitability are important issues when assessing its real impact. Under this premise, this paper presents an overview of the state of the art, including different approaches related to the detection, forecasting and propagation of vulnerabilities in software, focusing mainly on Machine Learning and Data Mining techniques.

Abstract— Breast cancer remains one of the leading causes of mortality among women worldwide, making early and accurate diagnosis crucial for improving patient outcomes. Traditional machine learning models often struggle to maintain high classification accuracy due to data variability and the complex nature of tumor characteristics. This study proposes an adaptive voting ensemble learning algorithm to enhance breast cancer classification performance. The ensemble integrates multiple classifiers—such as Decision Trees, Support Vector Machines, and K-Nearest Neighbors—by assigning dynamic weights based on each model's real-time performance. The algorithm is evaluated using the Wisconsin Breast Cancer Dataset (WBCD), and its effectiveness is measured against individual classifiers using metrics like accuracy, precision, recall, and F1-score. Results demonstrate that the adaptive ensemble significantly outperforms standalone models, offering a more robust and reliable approach for breast cancer prediction. This method shows promise for application in clinical decision support systems, contributing to more accurate diagnostics and better treatment planning. Keywords- Breast Cancer Classification, Adaptive Voting, Ensemble Learning, Machine Learning, Wisconsin Breast Cancer Dataset (WBCD), Decision Tree, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Medical Diagnosis, Predictive Modeling