The rapid increase in student enrollment at universities across the UAE has underscored the limitations of traditional academic advising methods, including long wait times and overburdened advisors. This paper presents a finetuned Academic Advisor Chatbot leveraging the “noushermes2” large language model (LLM) to deliver real-time, context-aware guidance tailored to the university's academic policies. By integrating institution-specific data, the chatbot addresses common student queries with high accuracy and clarity, as evidenced by BLEU scores improving from 0.45 to 0.78 and ROUGE-L scores reaching 0.85 for specific academic scenarios. A user feedback study involving 50 students revealed high satisfaction levels, with an average rating of 4.6/5 for accuracy, 4.7/5 for clarity, and 4.8/5 for time efficiency. The chatbot not only enhances the course registration experience but also reduces advisor workload, offering a scalable solution to meet the needs of a growing student population. Future enhancements will focus on multilingual support and expanded program-specific datasets to further improve usability and inclusivity.

Abstract Arrhythmia is a fatal cardiac clinical condition that risks the lives of millions every year. It has multiple classes with variable prevalence rates. Some rare arrhythmia classes are equally critical as common ones, yet are very hard to detect due to limited training samples. While several methods accurately detect Arrhythmia's multi-class, minority class accuracy remains low and these methods are resource-intensive. Therefore, most of the existing detection systems ignore minority classes in their classification or focus on binary classification. In this study, we introduce RL-ECGNet, a resource-efficient reinforcement learning-based optimization for multi-class arrhythmia detection, encompassing minority classes, through ECG signal analysis. RL-ECGNet uses raw ECG signals, processes them to extract the temporal ECG features, and utilizes Reinforcement Learning (RL) to optimize the training and network hyperparameters of the Deep Learning (DL) models while reducing resource consumption. For evaluation, four DL models, namely, MLP, CNN, LSTM, and GRU, are trained and optimized. Moreover, time and memory usage are minimized to optimize resource consumption. Throughout the evaluation of the four DL models, the proposed RL model achieved accuracies ranging from 88.45% to 96.41% for all 9 arrhythmia classes, including minority classes. In addition, the proposed RL method improved performance by a factor ranging from 1.28 to 1.39 in terms of accuracy. Moreover, the optimized DL models had reduced training time, as well as minimized memory usage. The proposed method achieved resource consumption reduction ranging from 1.36 to 1.925 times for training time, and from 1.179 to 1.815 times for memory usage.

Protecting and caring for water is one of the most critical environmental problems today. This research aims to design an intelligent system using machine learning models to improve water quality and predict whether it is safe to be used as drinking water. Several models of machine learning algorithms are compared to find the best model to be used for the accuracy of prediction of water quality. In this research, we compare Decision Tree, K-Nearest Neighbor, Support Vector Machine, Ransom Forest, and LightGBM models to get the best model for water potability prediction. Experimental results show that LightGBM model produced the best prediction accuracy of 99.74% on the experimental data.

Protecting and caring for water is one of the most critical environmental problems today. This research aims to design an intelligent system using machine learning models to improve water quality and predict whether it is safe to be used as drinking water. Several models of machine learning algorithms are compared to find the best model to be used for the accuracy of prediction of water quality. In this research, we compare Decision Tree, K-Nearest Neighbor, Support Vector Machine, Ransom Forest, and LightGBM models to get the best model for water potability prediction. Experimental results show that LightGBM model produced the best prediction accuracy of 99.74% on the experimental data.

This paper proposes a water leak detection framework for residential properties using machine learning. Water is an essential natural element, and water conservation is at the core of the global sustainable development goals set by the United Nations in 2015. While several research studies investigated water conservation at a larger scale, this study focus on residential water leakage and consumption. Several studies have reported that residential leakage can go unnoticed for a very long time, resulting in significant water waste. Sensor technologies and machine learning can help in the early detection of these leaks to minimize the wasted water at residential properties. The proposed leakage detection model was tested using a physical prototype, and experimental results show that the proposed model detects leakage and overuse with an overall accuracy of 87%. In addition, the proposed system proved efficacy in detecting almost all cases of the leak with a recall of 83%.

Due to the global outbreak of COVID-19, education was suddenly shifted to online platforms. While some institutes were ready for the transfer, others had to use available video conferencing tools for lecturing as a quick resolution. Zoom and Microsoft Teams arose to be the most used tools by institutes of higher education for online lecturing. Despite various reported challenges were attributed to lacking e-learning-related features in these platforms, other challenges were attributed to the usability of these platforms for lecturing activities. This study aims at evaluating the usability of Zoom and Teams for online lecturing activities based on a new set of heuristics focusing on cognitive styles. A thorough and systematic usability study is conducted based on the proposed heuristics. As a result of this evaluation, a set of critical, unmet user needs are identified and scored against the proposed heuristics. In addition, a set of guidelines are proposed to support informed selection of online lecturing platforms.

We address software engineering issues related to modeling game-based learning systems to support the kind of independent learning that takes place outside the classroom. After evaluating existing game-based learning (GBL) frameworks and identifying GBL characteristics, we elaborate and express formally a set of functional requirements that capture the many dimensions of GBL systems. From these requirements, we propose a GBL system structure that consists of an instructional subsystem, a control subsystem,, and a game subsystem. We use this model as a foundation to develop a fully operational GBL prototype intended for children to learn about the concepts in many microworlds. The game has a knowledge acquisition phase and a challenge phase. In the learning phase, the learner interacts with various objects and learn about their characteristics. In the challenge phase, the player is presented with several challenges related to objects he/she met. This phase is used implicitly to assess the impact of the first phase on the learner. The positive results of an experiment to assess the impact of this game on concept acquisition are discussed.

Information wikis and especially Wikipedia are attracting an increasing attention for informal learning. The nature of wikis enables learners to freely navigate the learning environment and independently construct knowledge without being forced to follow a predefined learning path in accordance with the constructivist learning theory. To the best of our knowledge, no effective personalized content recommendation approach has yet been defined to support informal learning from wikis. Therefore, we propose a personalized content recommendation framework that extrapolates topical navigation graphs from learners' free navigation and integrates them with fuzzy thesauri for automatic and adaptive personalized content recommendations to support informal learning in wikis. We design user studies and conceptual knowledge rubric to evaluate the impact of personalized recommendations on learning from Wikipedia. Results show that the proposed personalized content recommendation framework generates highly relevant recommendations. Evaluation of informal learning reveals that users who use Wikipedia with personalized recommendations can achieve higher scores on conceptual knowledge assessment compared to those who use Wikipedia without recommendations. Learners who use Wikipedia with personalized recommendations are able to utilize larger number of concepts and are able to make comparisons and state relations between concepts.

Wikis are attracting lots of attention for informal learning. The nature of wikis enables learners to freely navigate the learning environment and independently construct knowledge without being forced to follow a predefined learning path in accordance with the constructivist learning theory. Recommendation systems (RS) can provide useful content recommendations in different contexts. To our best knowledge, no effective personalized content recommendation approach has yet been defined to support informal learning in wikis. Therefore, we propose a personalized content recommendation framework to extrapolate topical navigation graphs from learners' free navigation and integrate them with fuzzy thesauri for automatic and adaptive personalized content recommendations to support informal learning in wikis.

We address the personalized/customized learning claim made by proponents of game-based learning (GBL), that is, GBL supports effectively learning personalization. Even though popular and fascinating, this claim remains unfounded. Indeed, in the literature on games, mentions of the multiplicity of exploration paths are put forth as examples giving players (learners) opportunities to tailor their learning experiences. This implicit assumption fails to demonstrate any kind of relationship between learning customization and playing games for learning. We explore some of the dimensions of the problem, describe a proposed framework, and use an example to demonstrate our approach.