Farman Ali Pirzado

Introduction: The rise of large language models (LLMs), such as ChatGPT, Codex, and Copilot, has transformed programming and education, providing powerful tools that enhance code generation, debugging, and conceptual understanding. Despite their potential benefits, concerns have emerged about their impact on students' learning behaviors, including over-reliance, trust issues, and engagement. Exploring how computer science students perceive and interact with LLMs is crucial for ensuring that these tools support, rather than hinder, independent learning and critical thinking. Without careful consideration, the widespread adoption of LLMs in programming education may lead to unintended consequences, including weakened problem-solving skills and reduced confidence in coding abilities. Considering this, the study explores students' perceptions of the role of LLMs in programming education, focusing on key factors such as perceived support, skepticism, engagement, dependence, and trust. Methods: To achieve this objective, we surveyed 300 undergraduate students from public and private institutions in Mexico. Overall perceptions were reported, and variable-wise comparisons—including gender and institution type were conducted using t-tests. One-way ANOVA was used to assess differences between semesters and levels of programming experience. Results: The results showed that, while students generally view LLMs as valuable tools, concerns persist about over-reliance, accuracy, and security. T-tests revealed no significant differences based on gender or institution type. ANOVA results indicated that programming experience, not academic semester, gender, or institute, was a key factor shaping perceptions. Students with 1-2 years of programming experience reported the highest levels of perceived support, engagement, and trust. Discussion: These findings underscore the importance of tailoring LLM integration strategies to students' prior experience to maximize benefits and mitigate risks in programming education.

Large language models (LLMs) have witnessed rapid adoption across academic and industrial domains due to their advanced capabilities, including code generation, natural language processing, and interactive support. Their integration into educational settings, particularly in programming has introduced both substantial benefits and emerging ethical challenges. The current study investigates students’ perceptions of LLM usage at two different public and private universities in Mexico, focusing on how ease of use influences their perceived effectiveness, confidence, future intention, and frequency of use in programming‐related tasks. A quantitative, survey‐based approach was adopted, with data collected from undergraduate and postgraduate students actively. The findings revealed that the ease of use plays a significant role in shaping students’ attitudes and behaviors. Path analysis using Partial Least Squares Structural Equation Modeling (PLS‐SEM) confirmed strong relationships between usability and perceived outcomes, suggesting that intuitive tool design can enhance confidence and increase regular usage. However, the study also highlights ethical concerns among students, including overreliance and potential academic dishonesty. These insights underscore the need for practical guidelines and institutional policies to ensure responsible and balanced integration of LLMs in academic environments. This research offers important implications for educators, developers, and policymakers working to align technological innovation with ethical learning practices.

Engineering education is based on experiential learning, but the problem is that in laboratory conditions, it is difficult to give feedback to the students in real time and personalize this feedback. The paper introduces the proposal of an innovative approach to the laboratories, called Adaptive Lab Mentor (ALM), which combines the technologies of Artificial Intelligence (AI), Internet of Things (IoT), and sensor technology to facilitate intelligent and customized laboratory setting. ALM is supported by a new real-time multimodal sensor fusion model in which a sensor-instrumented laboratory is used to record real-time electrical measurements (voltage and current) which are used in parallel with symbolic component measurements (target resistance) with a lightweight, dual-input Convolutional Neural Network (1D-CNN) running on an edge device. In this initial validation, visual context is presented as a symbolic target value, which establishes a pathway for the future integration of full computer vision. The architecture will enable monitoring of the student progress, making error diagnoses within a short time period, and provision of adaptive feedback based on information available in the context. To test this strategy, a high-fidelity model of an Ohm Laboratory was developed. LTspice was used to generate a huge amount of current and voltage time series of various circuit states. The trained model achieved 93.3% test accuracy and demonstrated that the proposed system could be applied. The ALM model, compared to the current Intelligent Tutoring Systems, is based on physical sensing and edge AI inference in real-time, as well as adaptive and safety-sensitive feedback throughout hands-on engineering demonstrations. The ALM framework serves as a blueprint for the new smart laboratory assistant.

The integration of artificial intelligence-based coding assistants (AI-CAs) into computing education introduces both opportunities and challenges, particularly for programming tasks. The effective adoption of these tools requires attention to students’ trust, learning motivation, and perception of AI-CAs’ reliability. Therefore, this study explores how undergraduate computer science students perceive and accept AI-CAs for programming tasks. The study analyzed computer science students’ perceptions of using AI-CAs, collecting responses from 311 students across different semesters in Mexico. This study extended the technology acceptance model (TAM) by incorporating six external factors: three from learning motivation (learning motivation, achievement goals, and subjective norms) and three trust-related factors (trust, doubt, and insecurity). The data analysis was performed using structural equation modeling (SEM). The results show that certain factors, such as learning interest, achievement goals, and trust, significantly enhance students’ perceived usefulness and ease of use, thereby increasing their intention to adopt AI-CAs for programming tasks. On the contrary, subjective norms, doubt, and insecurity showed limited influence on acceptance. Overall, this research validates the TAM in the context of AI-CAs in programming education and provides actionable insights for educators aiming to leverage AI tools effectively in computing education. It also offers fresh perspectives on adopting AI-CAs in programming education and reveals new perspectives on factors influencing students’ engagement with AI-CAs.

Autism Spectrum Disorder (ASD) encompasses various neurological disorders with symptoms varying by age, development, genetics, and other factors. Core symptoms include decreased pain sensitivity, difficulty sustaining eye contact, incorrect auditory responses, and social engagement issues. Diagnosing ASD poses challenges as signs can appear at early stages of life, leading to delayed diagnoses. Traditional diagnosis relies mainly on clinical observation, which is a subjective and time-consuming approach. However, AI-driven techniques, primarily those within machine learning and deep learning, are becoming increasingly prevalent for the efficient and objective detection and classification of ASD. In this work, we review and discuss the most relevant related literature between January 2016 and May 2024 by focusing on ASD detection or classification using diverse technologies, including magnetic resonance imaging, facial images, questionnaires, electroencephalogram, and eye tracking data. Our analysis encompasses works from major research repositories, including WoS, PubMed, Scopus, and IEEE. We discuss rehabilitation techniques, the structure of public and private datasets, and the challenges of automated ASD detection, classification, and therapy by highlighting emerging trends, gaps, and future research directions. Among the most interesting findings of this review are the relevance of questionnaires and genetics in the early detection of ASD, as well as the prevalence of datasets that are biased toward specific genders, ethnicities, or geographic locations, restricting their applicability. This document serves as a comprehensive resource for researchers, clinicians, and stakeholders, promoting a deeper understanding and advancement of AI applications in the evaluation and management of ASD.

The rapid integration of computational thinking (CT) into STEM education highlights its importance as a critical skill for problem-solving in the digital age, equipping students with the cognitive tools needed to address complex challenges systematically. This study evaluates CT skills among Engineering and Computer Science students using a multi-method approach by combining quantitative methods (CTT scores and CTS responses) with qualitative methods (thematic analysis of open-ended questions), integrating objective assessments, self-perception scales, and qualitative insights. The Computational Thinking Test (CTT) measures proficiency in core CT sub-competencies, abstraction, decomposition, algorithmic thinking, and pattern recognition through objective tests. The Computational Thinking Scale (CTS) captures students’ perceived CT skills. At the same time, open-ended questions elicit perspectives on the practical applications of CT in academic and professional contexts. Data from 196 students across two Mexican universities were analyzed through quantitative and thematic methods. The results show that students excel in pattern recognition and abstraction but face decomposition and algorithmic thinking challenges. Cross-sectional analyses were conducted between CTT, CTS and the open-ended part to compare CT skills across different demographic groups (e.g., age, gender, academic disciplines), showing clear differences based on age, gender, and academic disciplines, with Computer Science students performing better than engineering students. These findings highlight the importance of CT in preparing students for modern challenges and provide a foundation for improving teaching methods and integrating these skills into university programs.

As the integration of Artificial Intelligence tools, such as large language models (LLMs), into computing education increases, understanding their impact on students' learning becomes crucial. According to recent research, LLMs perform well when processing input in the English language. Still, they struggle when processing input in other languages or inputs containing non-English syntax or symbols, such as different languages and programming queries. Therefore, this study evaluates whether programming queries, particularly code generation queries in Spanish, a widely spoken language other than English, present challenges similar to those in code generation tasks compared to English queries. By doing this, this study aims to identify accuracy differences in the code generated by LLMs (Codex and Copilot) for English and Spanish input on a set of programming problems sourced from LeetCode. The study compares the performance of LLMs on three complexity levels of tasks, including basic, medium, and advanced code generation tasks. The results show that both Codex and Copilot show a significant decline in accuracy for Spanish as compared to English, particularly as task complexity increases from basic to advanced level. The Codex shows a significant decline in accuracy for Spanish inputs (85%) compared to English (92%). Similarly, Copilot shows a significant increase in accuracy for English inputs (93%) compared to Spanish (87%), with higher error rates across syntax, runtime, and logical errors in both. By comparing the results across multiple languages, the findings show that LLMs perform better on English-language inputs for code generation. Additionally, it demonstrated that Copilot also has superior adaptability and reliability in handling multilingual programming tasks compared to Codex. These results serve as a foundation and further emphasize the need for improvement in multilingual capabilities, as well as the language-dependent limitations of LLMs.

Rising breast cancer incidence and mortality represent significant global challenges for women. Deep learning has demonstrated superior diagnostic performance in breast cancer classification compared to human experts. However, most deep learning methods have relied on unimodal features, potentially limiting the performance of diagnostic models. Additionally, most studies conducted so far have used a single view of digital mammograms, which significantly reduces model performance due to limited overall perspective and generalizability. To address these limitations, we collected a multiview multimodal dataset, including digital mammograms four views two craniocaudal (CC), two mediolateral oblique (MLO) one for each breast, and textual data extracted from radiological reports. We propose a multimodal deep learning architecture for breast cancer classification, utilizing images (digital mammograms) and textual data (radiological reports) from our new in-house dataset. In addition, various augmentation techniques are applied to both imaging and textual data to enhance the training data size. In our investigation, we explored the performance of six state-of-the-art (SOTA) deep learning architectures: VGG16, VGG19, ResNet34, MobileNetV3, EfficientNetB7, and a vision transformer (ViT) as an imaging feature extractors. For textual feature extraction, we employed an artificial neural network (ANN). Afterwards, features were fused using an early fusion and late fusion strategy. The fused imaging and textual features were then inputted into an ANN classifier for breast cancer classification. We evaluated various feature extractors and an ANN classifier combinations, finding that VGG19 in association with ANN achieved the highest accuracy at 0.951. In terms of precision, again VGG19 and ANN combination surpassed other SOTA CNN and attention-based architectures, achieving a score of 0.95. The best sensitivity score of 0.893 was recorded by VGG16+ANN, followed by VGG19+ANN with 0.884. The highest F1 score of 0.922 was achieved by VGG19+ANN. VGG16+ANN achieved the best area under the curve (AUC) score of 0.929, closely followed by VGG19+ANN with a score of 0.915.