Abualsoud Hanani

This paper presents a toggle-control algorithm for multiple scan-chains Built-In Self-Test (BIST) in order to reduce peak and average power consumption during scanning-in test patterns. The proposed technique includes two phases. The first phase, which aims to reduce peak and average power, is implemented by adding a counter circuit to count the number of toggles of any test pattern and prohibits more toggles at a pre-specified limit. When this limit is reached, the remaining bits of the test pattern will be fixed at 0 or 1 to prevent more toggles. This phase although decreases the shift power (peak and average), it degrades the fault coverage (FC) of the Circuit Under Test (CUT). The second phase aims to improve the FC by reordering the cells within each scan-chain in such a way to maximize FC, and thus reducing test application time. Experimental results on ISCAS’89 and ITC’99 benchmark circuit’s show that using the proposed design can save on average about 34.54 % of average power consumption, 43.90 % of peak power consumption, while reducing in the same time the test length of about 4.05 %, which means huge power and energy savings without degrading test application time, and with negligible overhead in hardware area.

This paper introduces a novel approach for developing a learner model for school students by incorporating assessments from multiple sources, including students, parents, and teachers, in addition to exams. Traditional Intelligent Tutoring Systems (ITS) typically rely on a single source, such as exams, which often results in an incomplete picture of a student’s learning needs. To address this limitation, we developed a multi-source ITS that integrates feedback from various stakeholders to enhance the accuracy and personalization of the learner model. In a study involving 75 high school students, participants were divided into three groups: one receiving traditional classroom instruction, one using a conventional ITS, and one using our multi-source ITS. Results showed that the group using the multi-source system demonstrated the greatest improvement, with an average 38% increase in post-test scores, compared to 19% for the traditional instruction group and 26% for the conventional ITS group. Moreover, students using the multi-source ITS reported higher engagement and motivation, driven by their ability to interact with and adjust their learner model. These findings highlight the significant benefits of incorporating multi-source feedback into ITS to boost student performance, engagement, and motivation.

The market for mobile application development continues to thrive with billions of users and millions of apps. Collecting software requirements for mobile apps has to cope with this trend, so as for the software to compete in this crowded scene. Therefore, efforts to analyze mobile app reviews for requirements have shown a similar trend of increase. Among the billions of mobile users, there are hundreds of millions of Arabic-speaking users. According to our knowledge, this study would be one of the first studies in the field of mining mobile app reviews for the assistance of requirements engineering, to direct its focus on Arabic reviews. The main contribution of this study is to provide a framework for mining mobile app reviews in Arabic. A dataset of 7604 Arabic app reviews has been constructed and manually annotated by six experts. Each categorization aims at assisting one or more processes of software requirements engineering. Three configurations of deep neural networks, namely, CNN, LSTM, and BLSTM, were used to classify the app reviews into the considered categories of software requirements from the Arabic reviews. Furthermore, two word embeddings were utilized, on pre-trained models; Fasttext and Word2Vec, produced by this study. The sentimental analysis results show that the LSTM classifier with the Fasttext word embeddings gives the best F1-score, 79.17%. However, the BLSTM classifier with the fastText embeddings outperforms the other classifiers, with an F1-score of 69.83%, when used for identifying the sub-categories of the user perspective main category. The F1-score for classifying the sub-categories of the intention and topics with the LSTM and using fastText embeddings is 82.68% and 85,02%, respectively. These results outperform the other configurations of the classifiers and word embeddings. These results demonstrate the potential of our system to serve as a robust tool for automating software requirement extraction from Arabic app reviews, particularly in contexts where real-time user feedback is critical to agile development cycles.

Recent studies show that it is possible to detect technical dept automatically from source code comments intentionally created by developers, a phenomenon known as self-admitted technical debt. This study proposes a system by which a comment or commit is classified as one of five dept types, namely, requirement, design, defect, test, and documentation. In addition to the traditional term frequency-inverse document frequency (TF-IDF), several word embeddings methods produced by different pre-trained language models were used for feature extraction, such as Word2Vec, GolVe, bidirectional encoder representations from transformers (BERT), and FastText. The generated features were used to train a set of classifiers including naive Bayes (NB), random forest (RF), support vector machines (SVM), and two configurations of convolutional neural network (CNN). Two datasets were used to train and test the proposed systems. Our collected dataset (A-dataset) includes a total of 1,513 comments and commits manually labeled. Additionally, a dataset, consisting of 4,071 labeled comments, used in previous studies (M-dataset) was also used in this study. The RF classifier achieved an accuracy of 0.822 with A-dataset and 0.820 with the M-dataset. CNN with A-dataset achieved an accuracy of 0.838 using BERT features. With M-dataset, the CNN achieves an accuracy of 0.809 and 0.812 with BERT and Word2Vec, respectively.

Requirements engineering is the first phase in software development life cycle and it also plays one of the most important and critical roles. Requirement document mainly contains both functional requirements and non-functional requirements. Non-functional requirements are significant to describe the properties and constraints of the system. Early identification of Non-functional requirement has direct impact on the system architecture and initial design decision. Practically, non-functional requirements are extracted manually from the document. This makes it tedious, time-consuming task and prone to various errors. In this paper, we propose an automatic approach to identify and classify non-functional requirements using semantic and syntactic analysis with machine learning approaches from unconstrained documents. We used A dataset of public requirements documents (PURE) that consists of 79 unconstrained requirements documents in different forms. In our approach, features were extracted from the requirement sentences using four different natural language processing methods including statistical and state-of-the-art semantic analysis presented by Google word2vec and bidirectional encoder representations from transformers models. The adopted approach can efficiently classify non-functional requirements with an accuracy between 84% and 87% using statistical vectorization method and 88% to 92% using word embedding semantic methods. Furthermore, by fusing different models trained on different features, the accuracy improves by 2.4% compared with the best individual classifier.