The increasing digitization of modern society has led to a growing volume of digital information---such as emails, photos, videos, and electronic records---being presented as evidence in legal proceedings. However, traditional storage systems are costly, slow, and prone to server failures. To address these challenges, we propose a system leveraging blockchain and IPFS to store electronic evidence securely and reliably. Existing Layer 1 blockchain-based systems face three key limitations: (1) scalability issues with slow throughput and high costs under heavy traffic; (2) privacy risks due to public visibility of transaction data; and (3) impracticality of on-chain storage for large evidentiar... More >

Skin cancer has become a serious public health issue worldwide. The number of cases is rising due to higher UV exposure and changing lifestyle patterns. Early detection is the best way to save lives. However, many areas still lack specialized doctors and equipment needed for a quick diagnosis. In this paper, we develop and test an automated skin cancer detection method based on standard smartphone photography. It provides a practical solution for resource-limited settings by removing the need for expensive clinical dermoscopy equipment, making diagnosis accessible in areas with limited infrastructure. The system was trained and tested on PAD-UFES-20, which contains 2,298 clinical smartphone... More >

As data ecosystems become more diverse and time-critical, traditional monolithic ETL pipelines face challenges to meet the demands of modern data engineering workloads in terms of scalability, adaptability, and operational resilience. In this paper, we introduce an event-driven microservices approach to orchestrate and deploy AI-based ETL (ETL = Extraction, Transformation, and Loading) pipelines in a Kubernetes-managed environment that includes the following components: asynchronous orchestration using Apache Kafka, hybrid anomaly detection, adaptive schema inference, and predictive load balancing. The proposed architecture breaks the ETL processing into loosely coupled services, which can b... More >

The key feature of FinTech software systems is the ability to accurately assess risk in real time, making decisions on high-volume streams of information that are associated with very low latency and are robust to concept drift, and able to be updated without disrupting services. This paper addresses the problem of adaptive risk scoring using a reinforcement learning approach by modeling the risk evaluation problem as a continuous-action Markov Decision Process and continuously optimizing the policy via streaming transactional, behavioral events and outcome driven reward feedback. In addition to the learning algorithm, we also view ARL-CPO as a deployable software architecture that separates... More >

Automated classification of assessment questions according to Bloom’s taxonomy is increasingly used to support curriculum design and educational analytics. Many existing approaches rely heavily on instructional verbs as proxies for cognitive demand, despite longstanding concerns about their interpretive reliability. This paper adopts a data-centric perspective to examine why verb-centric Bloom-level classification remains fragile when applied to authentic multiple-choice question (MCQ) stems. The study is based on a custom, single-domain dataset of MCQ stems annotated according to the revised Bloom’s taxonomy, with intentional class imbalance preserved to reflect realistic assessment pra... More >

The concept of Return on Security Investment (RoSI) has evolved from a mere financial indicator into a comprehensive system for informed decision-making. Software-intensive organisations face mounting pressure to justify security expenditure in financially rigorous terms. Existing Return-on-Security-Investment (RoSI) models rely on deterministic approximations that ignore probability distributions over threats, temporal decay of vulnerability windows, and intangible cost categories. This paper presents a probabilistic RoSI framework grounded in the FAIR taxonomy that integrates: (i) expected-loss differentials with Bayesian updating; (ii) shift-left cost amplification across the software dev... More >

Non-Functional Requirements (NFRs) are quality-focused attributes of a system that impact functional components. There are 24 common NFR classes with some of the most used being performance, scalability, availability, reliability, and security. The implementations of these classes are ambiguous and describe the attributes on the behavior of software. Due to their nature, NFRs are typically realized through the specification and implementation of functional requirements. For traditional software systems, the NFR definitions are well known. However, in Machine Learning (ML), which has numerous applications across a multitude of domains, a current challenge is that traditional software non-func... More >

Electroluminescence imaging is widely used for detecting defects in solar cells. It reveals electrically active damage that remains invisible under conventional optical inspection. Most existing studies apply machine learning models to classify electroluminescence images and report performance mainly through accuracy scores. Inspection is often treated as an isolated prediction task, while physical defect mechanisms, sensing variability, representation bias, decision risk, and deployment constraints receive limited attention. As a result, strong benchmark results may not translate into reliable inspection outcomes in manufacturing environments. This paper presents a conceptual, non-systemati... More >