Generative artificial intelligence tools have become widely available in the past 3 years. Generative artificial intelligence is increasingly being adopted across diverse domains including the teaching of computer programming. In this study, four experiments were conducted to examine the effectiveness of generative AI in different roles within a course on programming. The students who used AI as a tutor reported higher motivation, programming self-efficacy and computational thinking of students, but the use of AI does not improve examination performance. When used as a problem solver, AI improves performance on programming tasks, but this improvement does not transfer to subsequent tasks completed without AI assistance. Additionally, AI can be utilized to evaluate and grade programming assignments. Although AI-generated grades are largely consistent with those assigned by a human expert, students exhibit low levels of trust in their accuracy. Furthermore, AI serves effectively as a design partner in complex programming tasks by proposing alternative solution strategies and facilitating discussion of trade-offs. In conclusion, generative artificial intelligence represents a promising technology with significant potential to enrich the teaching-learning process in programming education. However, its benefits can be maximized when applied to open-ended problems supported by well-structured and descriptive prompts.

Nielsen’s heuristics have served as a guideline for designing user interfaces and evaluating their usability for more than thirty years. Heuristic evaluation relies largely on qualitative analysis and there has been limited use of quantitative metrics in it so far. In this paper, we propose a new approach for heuristic evaluation based on Nielsen’s heuristics with an emphasis on quantitative metrics for assessing the usability of user interfaces. The approach is straightforward and suitable for use in the software industry. We used the approach to perform heuristic evaluation of four software products and five experts could identify 53 usability problems in the software products on an average. A quantitative analysis revealed that concurrence percentage, a measure of the extent of overlap among the usability problems identified by the different evaluators, was between 21 and 44

Human-computer interaction is a branch of computer science dedicated to the study of how human beings use computers and using the knowledge to develop hardware and software that are efficient and satisfying to users. The purpose of this paper was to develop an instructional approach in human-computer interaction in which students undertake a series of small projects requiring engineering and design skills. Accordingly, we designed three projects, (1) to design the user interface of a digital system taking into consideration performance-related and ergonomics-related parameters, (2) to design a new typeface for any specific application or user group, and (3) to test the usability of a digital system by objectively assessing measurable parameters, such as effectiveness and efficiency, and self-reported parameters, such as satisfaction and cognitive load. The projects involved interaction with end users and using state-of-the-art software tools. This approach was used to teach 104 undergraduate students during the autumn semester of 2024. An analysis showed that the projects provided the students an opportunity to collaborate with their peer and they performed better in examination.

Semiotics is the discipline that studies the signs and the cognitive process of meaning-making. As part of semiotics studies, the idea of semiospheres has been formulated, representing spheres of meaning that do not exist in isolation. This means that concepts can have multiple meanings. This paper shows these ideas applied to computer science and software engineering concepts as categories. We applied a survey to categorize different computing-related concepts, and we processed the answers of IT professionals from India and Chile. Findings show conceptual differences in the involved countries supporting the idea of technological semiospheres. We discuss the results from educational and disciplinary perspectives.

With rapid advancement of advanced driver assistance systems (ADAS) and autonomous vehicles, ensuring safe interaction between human drivers and ADAS system has become a critical challenge for accurately predicting driver behaviour. The study addresses the problem of developing robust systems, with limited raw features from smartphone inertial measurement unit (IMU) sensors. This study proposes a feature-driven approach to predict driver behaviour using smartphone IMU data. The objective is to enhance prediction accuracy by utilising two novel features namely, statistical deep features and cross-correlated features. These features capture more intrinsic patterns in driver behaviour and improve the performance. The proposed methodology was experimented on three benchmark datasets with machine learning models like random forest (RF) and extreme gradient boosting (XGBoost). Using cross-correlated features, accuracies of 99%, 100%, and 99% were obtained. This demonstrates that this approach outperforms existing methods, capturing detailed patterns and providing more reliable predictions in complex traffic.

Instant messaging apps are among the most widely used smartphone apps, but little research has been done on their usability. We conducted an experiment in which thirty volunteers used four common instant messaging apps for 2 to 3 weeks, and we observed their interaction with these apps in controlled conditions. We observed that effectiveness ranged between 69% and 96% and varied significantly (F = 19.551, p < 0.05) among the apps. Efficiency ranged between 50% and 70% and varied significantly (F = 5.447, p < 0.05). There was no major change in effectiveness and efficiency during the intervention period denoting low learnability. The satisfaction of the subjects varied significantly (F = 40.026, p < 0.05) for the apps. The affinity of three of the apps was above 75% denoting that three-fourth of the subjects who used them before the intervention planned to continue using them in the future. There was no strong influence of self-reported performance- and satisfaction related parameters on affinity.

Children are keen users of digital technologies, and we studied the interaction of 60 children aged 4 to 10 years with YouTube videos played on a smartphone. We found that the 9- and 10-year-old children liked watching videos on a smartphone, while the younger ones preferred TV shows. Most of the children aged 4 to 6 years made accidental touches on the screen of the smartphone while watching a video on it, and such accidental touches sometimes paused or closed the video. Most children aged 7 years and more could use the voice search feature of YouTube to look for a specified video, use the different widgets of the website, and even skip through advertisement. The average attention span of the children while watching videos on a smartphone increased from 3 to 9 min between 4 and 10 years of age. We recommend videos meant for children have content to facilitate informal learning and nurture creativity in them, be of appropriate length, and avoid small objects and crowded scenes. Video streaming websites should have a “child mode” with a minimalistic user interface and support for voice-based interaction.

The proposed research presents a comprehensive method for diagnosing and detecting mispronunciations, using a Variational Autoencoder (VAE) to improve performance. The proposed approach combines several feature variables and modalities for a more efficient analysis. Using VAE as an audio encoder, audio data representations are captured to learn a compact and useful latent space representation. Bi-directional Long Short-Term Memory (Bi-LSTM) network is used as a phoneme encoder to extract phonetic information from the input data. Transformer-based decoder is incorporated to decode the learnt representations and provide sequences associated with speech pattern pronunciation. To improve the model’s comprehension of the speaker’s pronunciation material, Multi-Modal Fusion incorporates phoneme information both before and during the different phases of VAE. The implementation of a secondary decoding mechanism is part of the decoding process. This method entails returning the decoded sequence to the decoder for a further round of decoding. This improves mispronunciation diagnosis and identification by overcoming the difficulty of incomplete prior knowledge during the initial decoding stage. The L2 Arctic English dataset is used for the experimental evaluation of the model. Significant improvements are shown as compared to the baseline model which used Squeezeformer as an encoder (Guo et al. in “Multi-feature and multi-modal mispronunciation detection and diagnosis method based on the squeezeformer encoder”. IEEE Access (99):1–1). The accuracy increased slightly from 96.19 to 96.27

Smartphones are ubiquitous platforms and the most prevalent digital media devices use by university students for various activities. The lockdowns due to COVID-19 and online education are making students use smartphones frequently and for longer durations leading them to feel nomophobic. The objective of this study is to identify the prevalence of nomophobia among university students during lockdowns in India using unsupervised machine learning. This study is conducted on 111 undergraduate students (mean age: 19.93 years, S.D.: 0.69) in an Indian university volunteered. Their perception about their smartphone using a questionnaire and their smartphone usage pattern using a purpose-built android application has been studied. It has been found that 12.6% of the students were suffering from severe nomophobia and another 78.4% of them were suffering from moderate nomophobia. The average daily smartphone usage and the average daily nighttime smartphone usage of the students were 8.17 hours (S.D.: 2.73) and 5.38 hours (S.D.: 2.03), respectively. The average daily phone unlock frequency was 132.64 (S.D.: 64.01). Nomophobic behavior was found be associated with daily smartphone usage and daily nighttime smartphone usage of the students using Pearson's correlation. An analysis of smartphone dependence and usage pattern using unsupervised machine learning revealed that 18.9% of the students were at a high risk of smartphone addiction.

Digital computers were invented in the 1940s. They are sophisticated and versatile machines whose functioning is grounded in elaborate theory. Advances in theory and the availability of computers helped computer science to develop as an academic discipline, and university departments for the same started coming up in the 1960s. Computer science covers all phenomenon related to computers and consists primarily of man-made laws governing building, programming, and using computers. Computational thinking is a way of thinking influenced by computers and computer science. There are two schools of thought on computational thinking. The first school sees computational thinking as the use of computers to explore the world, while the other sees computational thinking as the application of concepts from computer science to solve real-world problems. Scholars typically agree that computational thinking has four essential components, viz., abstraction, decomposition, algorithm design, and generalization. Computational thinking is often feted by computer scientists as a useful skill that can be used by anybody anywhere. However, it is necessary to find out ways for successfully using computational thinking in domains other than computer science before it can be declared a universal skill.

Wearable sensor‐based devices like actigraphs collect motor activity data which provide objective measures of physical activity. This research puts forward a novel methodology for assessment of objective sleep quality using actigraph recordings of motor activity. High level features of sequential motor activity data are extracted using Long‐Short Term Memory (LSTM) model which are then paired with a significant statistical feature namely, zero percent which describes the percentage of events with zero activity over a series. Overlapping sliding window is used to input sequences into LSTM to capture superior features in activity recordings. The predictive ability of the combined feature vector is evaluated using support vector machine (SVM) classifier. This hybrid LSTM‐SVM framework is validated on a benchmark dataset namely, the MESA Actigraphy dataset and achieves an accuracy of 85.62% for sleep quality prediction. Effectiveness of overlapping sliding window and statistical feature are evaluated, and their significance is validated. It is validated that the concept of overlapping sliding window improves the performance accuracy by 3.51% and the use of discriminative statistical feature improves sleep quality prediction task by 2.95%. Comparison with state of the art validates that this is the first study using objective sleep quality indicator for assessment of sleep quality via actigraph‐based motor activity data.

Many high school students find algebra difficult because it requires comprehending abstract concepts, performing symbol and sign manipulation, and solving multi-step problems. Graphing calculators can visualize algebraic expressions and equations, and make learning more interactive. We conducted an experiment in which Desmos online graphing calculator was used to augment the algebra lessons of tenth-standard students. The students were taught about polynomials, linear and quadratic equations, and arithmetic progressions during an intervention period lasting seven weeks. We found that properly integrating the graphing calculator with classroom sessions and homework helped the students to perform better in the post-test and also retain the concepts. The students found the graphing calculator to be easy to use and helpful in learning, which increased their interest in algebra and encouraged them to practice algebra regularly.

People with significant intellectual disabilities frequently struggle to articulate their stress, which makes timely carer reactions difficult. Stress, a frequent problem in modern life, is often overlooked, emphasising the importance of early detection. Smartphones, which are continually used, generate a variety of behavioural data, making them perfect for stress detection. This paper offers an autonomous real-time stress detection system that use wearable sensors to capture physiological signals and machine learning to detect stress-related changes. Our study uses data from smartphone sensors, including linear acceleration, gravity, gyroscope, and touchscreen sensors, to detect stress through soft keyboard typing behaviours. We capture nuances in phone interactions under stress by extracting statistical time domain features from raw sensor data, such as mean, variance, standard deviation, and energy. Using feature selection techniques, we reduced the dataset to the most significant features, minimizing model complexity and improving accuracy. We applied advanced machine learning methods like bagging, boosting, and stacking classifiers. The AdaBoost classifier achieved a 94.21% accuracy, a 5.58% improvement using the selected feature subset compared to the full feature set. The optimized model also significantly reduced the model-building time to 1.73 seconds, demonstrating the efficiency and practicality of our approach for real-time stress detection.

The integration of different Mobile Edge Computing (MEC) applications has significantly enhanced the realm of security and surveillance, with Human Activity Recognition (HAR) standing out as a crucial application. The diverse sensors found in smartphones have made it convenient for monitoring applications to gather and analyze data, rendering them valuable for HAR purposes. Moreover, MEC can be employed to automate surveillance, allowing intelligent monitoring of restricted areas to identify and respond to unwanted or suspicious activities. This research develops a system using motion sensors in smartphones to identify unusual human activities. People's smartphones were employed to monitor both suspicious and regular activities. Information was collected for various actions categorized as either suspicious or regular. When a person performs a certain action, the smartphone records a series of sensory data, analyse important patterns from the basic data, and then determines what the person is doing by combining information from different sensors. To prepare the data, information from different sensors was aligned to a shared timeline. In this study, we used a sliding window approach on synchronized data to feed sequences into LSTM and CNN models. These models, which include initial layers of LSTM and CNN, automatically find important patterns in the order of human activities. We combined SVM with the features extracted by the shallow Neural Network to make a mixed model that predicts suspicious activities. Lastly, we compared LSTM, CNN, and our new shallow mixed neural network using a new real-time dataset. The mixed model of CNN and SVM achieved an accuracy of 94.43%. Additionally, the sliding window method's effectiveness was confirmed with a 4.28% improvement in accuracy.

Wearable devices are equipped with inertial sensors that can collect motion data and provide objective measures of a person's physical activity. Smartphones are increasingly used to monitor users' activity, enabling accurate qualitative and quantitative measurement. In human activity patterns, aberration refers to any action that deviates from the normal or expected course of action. In the case of physical activity, aberrant activity must be continuously monitored and reported promptly in real-time scenarios. This study employs monitoring real-time physical activity features through the utilization of smartphone inertial sensors, to distinguish between aberrant and non-aberrant activity classes. Data from multiple sensors including accelerometer, gyroscope, magnetometer, and others, are collected from five participants' smartphones and synchronized to monitor activity. To obtain optimal set of statistical features, three meta-heuristic approaches—namely elephant search, wolf search, and cuckoo search—are utilized, and in combination with a correlation attribute, serve as feature filtering methods for feature selection. A new optimal feature set is derived from this process, resulting in a reduction in cardinality by approximately 43

The COVID-19 pandemic forced organizations and employees to switch to working from home. We conducted a survey to understand the opinion of people on working from home, and if they would like working from home after the pandemic is over. We found that the respondents, most (92%) of whom were in IT-related professions, appreciated the comfort and flexibility afforded by working from home, but also had difficulties because of lack of facilities at home, low motivation and reduced synchronous communication, collaboration and teamwork. The possibility of IT-related activities (t= 2.214), flexibility (t= 2.603) and effective communication (t=2.012) made the respondents feel that working from home can lead to positive changes in society (P<0.05). The shift to working from home allowed organizations to continue with their business processes during the pandemic and lockdowns, and the experience was mostly positive for employees. The shift to working from home also brought a much needed change in the work culture in many organizations. organizations may allow their employees to work from home as a norm rather than as an exception after the pandemic but will require redesigning their business processes.

This research puts forward a methodology for depression assessment using actigraph recordings of motor activity. High level features of motor activity are extracted using Long-Short Term Memory (LSTM) which are paired with statistical features to deliver valuable digital biomarkers. Overlapping sliding window is used to input sequences into LSTM to capture superior features in activity recordings. The predictive ability of these digital biomarkers is evaluated using Support Vector Machine (SVM). The hybrid framework is validated on benchmark, Depresjon dataset and achieves accuracy of 95.57

Smartphones have been owned and used ubiquitously in all facets of society utilized for a wide number of tasks such as calling and messaging, social media, surfing as well as for entertainment.Spending a large amount of time on smartphone might lead to a dependence on it for a variety of purposes.This study uses objective measures of real time smartphone usage features to assess smartphone addiction.A purpose built android application to collect real time smartphone usage has been developed and linear classification models namely Support Vector Machine and Logistic Regression are used to predict smartphone addiction among university students.Furthermore, correlation and information gain measures are used to identify most vital features of smartphone usage which contribute maximum in assessment of smartphone addiction.It has been observed that both the linear models give worthy performance with more than 80% of accuracy.Also, the most important technical features impacting smartphone addiction are longest session spent for entertainment, total time used for communication, longest session spent for communication, longest session spent for work, total time used for entertainment, longest session for news and surfing, and data usage in other activities.

Schools operated mostly in online mode in the last two years because of the COVID-19 pandemic and many teachers are now using digital tools in physical classrooms after the reopening of schools. We conducted an experiment on how smartphones can be used to reinforce biology lessons of seventh standard students (mean age = 12.22 years) in a private school in India. Students of four sections were taught about respiration in human beings and other organisms following four different approaches. The first section of students was taught without using any digital tool, while the second, third and fourth sections were taught using YouTube videos, a tutorial app that provided study materials and quizzes for self-assessment, and an interactive app that displayed a three-dimensional model of an organ system from which students could learn interactively, respectively. The intervention lasted one month and its effectiveness was assessed by a post-test and collecting feedback from the students. The students of the third and fourth sections exhibited better lower-order thinking skills in the post-test, while the students of the second and fourth sections exhibited better higher-order thinking skills. Further, the students felt that they could easily use the tool they were taught with and learn details of the respiratory system from it. The students of the second and fourth sections typically found the diagrams and animation in the YouTube videos and the interactive app, respectively to be useful, while the students of the third section did not find the diagrams in the tutorial app to be of much use. We concluded that smartphone apps are attractive to children and their long-term teacher-guided use can help children to acquire useful knowledge and skills.

Computational thinking is the process of solving a problem in a way that can be readily automated by a computer. Decomposition, pattern recognition, abstraction, and algorithm design are four major components of computational thinking. We conducted an experiment to study the effectiveness of smartphone apps for teaching computational thinking to 7-year-old children. The first experimental group was taught computational thinking using four apps, one for each component. The second experimental group was taught computational thinking using the chalk and board approach. The control group was not imparted any lesson in computational thinking. The intervention period lasted 8 weeks and was followed by a posttest and a retention test conducted after another 2 weeks. The children showed interest in learning computational thinking and could solve problems that are inspired by real-world engineering problems during the intervention period. They could analyze and debug (77%) their own work. The performance of the children in the two experiment groups was significantly better than those in the control group in the posttest and the retention test (F = 26.470, p < .05). This means that 7-year-old children can learn computational thinking from suitable mediums. No significant difference was observed in the performance of the children in the posttest and the retention test (p > .05), denoting that children can retain the concepts of computational thinking. Including computational thinking lessons at the K-12 level can encourage children to enroll in engineering programs in the future. We recommend standardization of content and development of age-specific apps for teaching computational thinking.