Requirement management is a fundamental process in software engineering. However, the traditional way of requirement management can result in unclear, ambiguous, inconsistent, and redundant requirements that can cause rework, cost and schedule overruns. In this paper, we present the design and development of an artificial intelligence-driven requirement management system that utilizes natural language processing techniques to enhance the clarity, completeness, and consistency of software requirements. Our proposed system provides support to several requirement management processes such as defect detection, conflict detection, and impact analysis. We conducted a survey to understand the current challenges and problems faced during requirement management. Then, we designed and developed the proposed system. The paper presents an Agile software development methodology to develop the proposed system that can accommodate continuous improvement and feedback from users. Our testing demonstrate that our proposed system can improve the clarity and traceability of software requirements throughout the software development life cycle.

Despite the availability of Artificial Intelligence applications across different domains, persistent usability challenges have hindered their effectiveness and user adoption. This systematic review investigates the role of interaction design in enhancing the usability and user experience of Artificial Intelligence applications. This analysis yielded three key findings. First, integration & interaction, learning, and communication are the primary Artificial Intelligence domains that leverage interaction design research to ensure user-centric and efficient applications. Second, Artificial Intelligence techniques such as hybridization, searching, and optimization are widely adopted to address multifaceted challenges. Finally, mixed methods have emerged as the most versatile user experience evaluation approach, integrating quantitative and qualitative techniques to provide comprehensive insights into Artificial Intelligence applications. These results offer a holistic framework for Artificial Intelligence practitioners and designers to develop intuitive and robust applications while addressing gaps in the field.

In times where working remotely and online work environments are on the rise; the traditional employee training method struggles to stay motivated and engaged. Gamification has emerged as a widely used motivational approach that can be tailored to user preferences to improve user engagement. While various machine learning techniques have been applied to gamification, the use of reinforcement learning for real-time adaptive personalization in employee training contexts remains comparatively underutilized. Therefore, this study presents a Reinforcement Learning–Based Dynamic Adaptive Gamification (DAG) method designed to enhance online employee training through real-time adaptation and personalization. The proposed approach leverages data-driven decision making, iterative feedback mechanisms, and personalization algorithms to optimize the training experience. The methodology comprises three key phases: (A)System design, (B)Implementation consisting the Adaptation and Personalization processes such as game interaction, ongoing feedback and (C)Evaluation. The results indicate the effectiveness of the DAG method in improving user engagement and increasing user retention in online training environment.

Mobile gaming is an area of digital entertainment that provides platforms for social engagement, education, fun, and cognitive development. The success of mobile games is influenced by interaction design, which encompasses the user interface, usability, and user experience. This systematic review examines recent developments in interaction design approaches, usability principles, and factors affecting the uptake of interaction design in mobile gaming. Thirty primary studies, published between 2020 and 2024, were examined. In addition, the findings show that user-centered design is the most common approach, which emphasizes user participation throughout the design process. User control and freedom are highlighted as critical usability principles that emphasize the importance of users in controlling their actions and learning at their own pace. Technological factors are important for the adoption of an interaction design in mobile gaming.

In recent years, there has been a notable increase in the utilization of gamification for online training. Following the “one-fits-all” approach to designing a gamification experience for all participating users can be a significant disadvantage as all users are given the same experience. Previous approaches commonly used the static player profile that was obtained at the very beginning or initial stage of the user experience and presented the user with the gamification experience according to the fixed player type. This motivates the need for an adaptive gamification environment in online training. A dynamic adaptive gamification framework is proposed to introduce a gamification framework consisting of a method to correlate player types and game elements. The correlation is aimed at identifying the evolvement of players through interaction with game elements and observing how these player profiles are motivated to change over time. The study also presents an evaluation method to measure motivation in adoption and engagement in online training to reduce boredom and improve the overall user experience of the player.

The iterative nature of agile requirements engineering often requires considerable time and effort. Research on automation processes with artificial intelligence for agile requirements engineering practices has gained popularity in recent years. The objective of this study is to conduct a systematic literature review and report the current state of artificial intelligence integration in agile requirements engineering. 24 papers were selected for the analysis. From the papers included in the study, it was found that the most popular requirements engineering activity currently being researched with artificial intelligence is requirement analysis and documentation, with natural language processing tools combined with machine learning as the most used artificial intelligence tool. The levels of automation were still mostly at level 3 and level 4. Similar to related works, this indicates that the implementation of agile requirement engineering largely depends on humans.

Non-functional requirements (NFRs) are often poorly documented and inadequately integrated into the software development process resulting in discrepancies between expected and actual system quality. This gap leads to a number of issues like requirement ambiguity, ineffective stakeholder communication and inadequate validation which severely impact project outcomes. This systematic review addresses these challenges by investigating the state of NFRs management in requirements engineering. The review identifies predominant techniques such as automated tools and document analysis for eliciting NFRs while also highlighting current challenges and effective validation practices, including inspections, reviews and user feedback. Future research should develop comprehensive frameworks combining technology with human-centered approaches to address these challenges.

The present study aims to build a hybrid convolutional neural network and transformer UNet-based model, Trans-Swin-UNet, to segment ischemic lesions of the plain computed tomography (CT) image. The model architecture is built based on TransUnet and has four main improvements. First, replace the decoder of TransUNet with a Swin transformer; second, add a Max Attention module into the skip connection; third, design a comprehensive loss function; and last, speed up the segmentation performance. The present study designs two experiments to evaluate the performance of the built model using both the self-collected and public plain CT image datasets. The model optimization experiment evaluates the improvements of Trans-Swin-UNet over TransUnet. The experimental results show that each improvement of the built model can achieve a better performance than TransUNet in terms of dice similarity coefficient (DSC), Jaccard coefficient (JAC), and accuracy (ACC). The comparison experiment compares the built model with four existing UNet-based models. The experimental results show that the built model had a DSC of 0.72±0.01, a JAC of 0.78±0.04, an ACC of 0.75±0.03 using the self-collected plain CT image dataset and a DSC of 0.73±0.02, a JAC of 0.79±0.03, an ACC of 0.76±0.02 using the public plain CT image dataset, achieving the best segmentation performance among five UNet-based neural network models. The two experimental results conclude that the built model could accurately segment ischemic lesions of the plain CT image. The limitations and future work of this study are also discussed.

A user story is commonly applied in requirement elicitation, particularly in agile software development. User story is typically composed in semi-formal natural language, and often follow a predefined template. The user story is used to elicit requirements from the users' perspective, emphasizing who requires the system, what they expect from it, and why it is important. This study aims to acquire a comprehensive understanding of user stories in requirement elicitation. To achieve this aim, this systematic review merged an electronic search of four databases related to computer science. 40 papers were chosen and examined. The majority of selected papers were published through conference channels which comprising 75% of total publications. This study identified 24 problems in user stories related to requirements elicitation, with ambiguity or vagueness being the most frequently occurring problem reported 18 times, followed by incompleteness reported 11 times. Finally, the model approach was the most popular approach reported in the research paper, accounting for 30% of the total approaches reported.

The study aims to develop a virtual rehabilitation system to assist upper limb motor training for older post-stroke patients. The system contains data-driven virtual exergames simulating the task-oriented training; receives the rehabilitation prescription and the online data collected from the multi-mode hand controller and the depth camera; assesses online patient's performance which in turn updates the data of virtual exergames to adapt to the patient's training progress. Its innovation is providing precision rehabilitation via gaining the personalized learning experience to improve the adherence to and effectiveness of virtual rehabilitation.

With the rapid development of Web technologies, online learning has become mainstream. People acquire for new skills or knowledge through the Internet in formal or informal ways. As compared to the well-developed curriculum-based Learning Management Systems (LMSs), more people prefer the informal style-learning that can be customized to meet different learning habits or interests. They prefer engage with resources that are short or casual, but could give immediate solutions. In informal online learning, people acquire learning resources from open-source platforms. As the open-source learning resources grow, the searches become complex and nearly impossible to identify desired content if just have a glance at the learning resource title. This paper described the full architecture of “MVR-RCM” to help learners acquire microlearning video resources efficiently in an informal online learning environment. Instead of manually browsing through the Internet for learning resources, an automated approach named “MVR-RCM” has been proposed to deliver microlearning videos using a content-based recommender system. In “MVR-RCM”, key influence factors like learning interest and video's category are adopted to suggest relevant microlearning video resources to learners.

Digital platforms have been widely adopted during Covid19 pandemic, but individuals and organizations have encountered various challenges with these digital platforms. Metaverse is a new Internet evolution that addresses the challenges of supporting humans in seamlessly moving between digital spaces and the physical world. Many bibliometric studies have been published on Metaverse, but previous findings on this topic are inconsistent and contradictory. Therefore, we explored trends in Metaverse research through bibliometric analysis. An online search was conducted on 16 February 2023. A total of 706 journal documents were obtained from the Scopus database. The results revealed that China was the most productive country with three China productive institutions, two China productive authors, and four China funding sponsors. This bibliometric analysis results confirmed that the rate of Metaverse research has sharply increased. The top nine most cited documents were classified as secondary research. Taking a socio-technical design approach, the Metaverse research areas identified by these review articles can be grouped into four main levels of computing: mechanical, informational, psychological, and social. The limitations of this bibliometric study were also identified.

A new design of web-based translation, also known as easyTranslate, is proposed for non-native English-speaking students to translate a paragraph of academic text. However, there is a need to determine whether this new web-based translation is usable or not for students during the Covid-19 pandemic. Thus, this paper describes a remote moderated usability testing of a web-based translation. Five undergraduate students took part in the usability testing through an online video conferencing tool. The results showed that all students successfully completed the given tasks and the new web-based translation is considered usable. The advantages and challenges of the remote moderated usability testing are discussed.

Software Requirements Prioritization (SRP) is one of the crucial processes in software requirements engineering. It presents a challenging task to decide among the pool of requirements and the variance of the stakeholder’s needs in prioritizing requirements. Semi-automated requirements prioritization is implemented in both manual and automatic processes. When prioritizing requirements, these aspects such as importance, time, cost and risk, should be taken into account. The emergence of machine learning is advancing to improve and automate the SRP process whereby decision making can be performed with minimal human intervention. Incorporating machine learning approaches in prioritization techniques can be implemented in the ranking process and classifying the priority group of the software requirements. A Semi-Automated Requirements Prioritization framework (SARiP), which implements semi-automatic process in requirements prioritization is proposed. SARiP concentrates on the areas related to prediction of requirements priority group and ranks requirements using classification tree and ranking algorithm. SARiP has been successfully evaluated in the government sector domain by the i-Tegur team from the Department of Information Technology, Ministry of Housing and Local Government of Malaysia (KPKT). 80% of the participants agreed that SARiP is extremely likely to help the participants in prioritizing the requirements for their projects. All participants agreed that SARiP is reliable and useful. Recording the requirements and results for the prioritization will be considered for future work and traceability function will be included to trace the requirements changes.

Online training is expected to increase retention of information and be less time-consuming. This leads to a motivation to identify a more effective content delivery for online training. Microlearning indicates that bite-sized content is delivered in short fragments that can fit into anyone’s hectic schedule. However, the perspective of microlearning and its content design is still indefinite. It is challenging to design content for training that optimizes microlearning’s characteristics. The purpose of this research is to identify the perspective towards microlearning and the significance of the design of micro-sized content for online training for employees. This study investigated two questions which are how to design an effective micro-sized content for enhancing employees learning opportunities and the type of topics which are relevant to learning. The study was carried out with employees from education industries and training service providers. Data was collected through a survey and focus group interview. The study recognised that employees are primarily familiar only with video-based microlearning content and they have inadequate knowledge on the application of other microlearning elements for content design. One of the most common microlearning elements – Video, that is between 5-7 minutes in length, is considered to be the most applicable element in microlearning. In conclusion, the perspectives concerning the challenges in designing microlearning content were discussed. The study also proposed 2 different architectures with different objectives for the overall microlearning content design based on the employee’s experiences and perspective.

In the big-data era, massive Open Educational Resources (OERs) can be obtained from the Internet regardless of location or time constraints. Researchers have discussed microlearning as a service to improve learning effectiveness. However, the emergence of OERs leads to the challenge of searching for appropriate and relevant microlearning resources. In this paper, an automated video classification approach named “MVR-CLS” is proposed to organize and classify microlearning resources, so that the learners can browse for learning resources in a manageable way. Speech-To-Text data mining technique is applied to transcribe a learning video and to further analyze the video content. A 3-tier learning category structure is proposed to organize a collection of microlearning videos into appropriate learning categories. “MVR-CLS” has shown the capability to classify the microlearning videos into a finer-grained learning category as compared to the existing work. To evaluate the accuracy of the proposed approach, the classification result is validated against to the metadata of the OERs. The classification result can promote better fit of learners’ interests for content recommendations and thus enhancing the recommendation accuracy in future work.

The evolution of digital technologies is leading the world towards the direction of the information explosion. It gradually increases the difficulty for the people to find appropriate content to learn. It has becoming a norm whereby people often use their fragmented spare time for learning. It leads to the motivation to look for a solution to boost up the learning effectiveness. Microlearning serves as a service to generate and deliver microlearning resources to the learners. However, it is also challenging to convey the microlearning resources to each learner based on different learning needs. In this paper, a personalized microlearning framework named “Unique-Learn” is proposed. It possesses the intelligence to identify the real learning needs of a learner based on the contextual information, then conveys the appropriate microlearning videos to the learner from time to time. The proposed implementation plan details how the “Unique-Learn” will be used in a workplace environment for the employee’s training and development purpose.

In Malaysia, traditional teaching approach is still the most common approach used in the tertiary mathematics education.Under the traditional teaching approach, students learn mathematics through rote-learning and being "spoon-fed".Traditional teaching approach may lead to students have loss of interest in learning mathematics and consequently, students seriously do not reach in-depth understanding of mathematics.This shows that students do not learn mathematics effectively and efficiently.The motivation of this research is to find out student's learning effectiveness and efficiency in learning the tertiary mathematics course under a designed blended learning approach that incorporates the core-and-spoke model.A preliminary statistical analysis is conducted at early stage to give a better understanding on the collected data.The analysis of student's learning effectiveness is carried out based on the student's learning outcome achievement and learning satisfaction.Technical efficiency in data envelopment analysis (DEA) is employed to evaluate the student's learning efficiency.The results show that the students not only have more interest in learning the tertiary mathematics, but also they are able to learn in a more effective and efficient way under the designed blended learning approach.

Requirements elicitation is one of the most essential activities in requirements engineering. The last twenty years have seen a growing trend towards Internet of Things and these new applications imposed a challenge for requirements engineers to elicit requirements. The aim of this paper is to identify and present the current trends of elicitation techniques that have been applied in Internet of Things application. To achieve this aim, an electronic search on two computing-related databases was included in this systematic review. After a thorough scanned independently done by the writers, 12 selected publications were examined. The findings of this systematic review revealed that (1) home and public spaces were prevalent domains for Internet of Things applications, (2) interviews and prototypes were most frequently used elicitation techniques, and (3) stakeholders were the common requirements sources.

An analogical-based approach towards meaning preservation by transferring the source language meaning structure to the target language in the recombination process of an English to Malay Example-based Machine Translation system is presented. The meaning structure is built on top of the current synchronized translation examples pair representation with the incorporation of a layer of semantic annotation at the structural level. This meaning structure provides a consistent medium allowing the derivation of translation hypothesis using analogical-based approach throughout the automated translation process. The complexity of the structural transformation in the final recombination process is relaxed with this analogical-based derivation approach. The preliminary experiment demonstrates that the English to Malay automated translation is improved using the analogical-based approach. The best evaluation score is obtained using the Bilingual Evaluation Understudy metric, showing improvement of 37.06%.