Programming Tasks Research Articles

Metacognitive control during problem solving at early ages in programming tasks using a floor robot

This research, following a sequential mixed-methods design, delves into metacognitive control in problem solving among 5- to 6-year-olds, using two floor-robot environments. In an initial qualitative phase, 82 pupils participated in tasks in which they directed a floor robot to one of two targets, with the closer target requiring more cognitive effort due to the turns involved. The results of this phase revealed that younger students often rationalised their decisions based on reasons unrelated to the difficulty of the task, highlighting limitations in children’s language and abstract thinking skills and leading to the need for a second quantitative study. In this subsequent stage, involving 117 students, a computerised floor-robot simulator was used. The simulator executed the students’ planned movements and provided feedback on their validity. Each participant had three attempts per problem, with the option to change their target each time. The simulator stored the information pertaining to the chosen resolution path, design of the plan, and re-evaluation of decision making based on the results and feedback received. This study aims to describe the criteria upon which students base their metacognitive control processes in decision making within problem-solving programming tasks. Additionally, through a comparative analysis focusing on age and gender, this research aims to assess the relationship between metacognitive processes and success in problem-solving programming tasks.

Health literacy pedagogies and assessment embedded in a school-based health education program

AbstractStudies have focused on the development of teachers’ Health Literacy (HL) pedagogies but have yet to focus on building teachers’ assessment capacity. The aims of this research study were three-fold: 1) professionally develop teachers to enhance the HL planned in a Health and Physical Education (HPE) program and assessment task; 2) validate a previously developed curricular model of HL assessment (rubric) using student work samples; 3) measure students HL learning using an adaptation of the validated rubric. Researchers used a descriptive single school case study, using a mixed methods approach in an Independent, coeducational early learning, primary and secondary day school in Sydney, Australia to work closely with participants to review the pedagogical knowledge, learning activities and assessment within a HPE program using Nutbeam’s three level HL hierarchy. Using Messick’s six aspects of construct validity testing of the rubric was completed using 20 sample assessment tasks, as well as receiving feedback on the adaptability and feasibility of the rubric. The validated rubric was then adapted to the assessment task to measure students’ HL. Results demonstrated that professional development had a positive impact on the development of planned HL pedagogies as shown by a 17% increase in planned critical HL learning activities. Findings suggest that the validated rubric is a feasible and adaptable tool that teachers can use to measure students’ HL. This is the first study to validate a curricular model of measuring HL. Future research needs to develop teachers’ HL capabilities and assessment literacy.

Agriculture students’ use of generative artificial intelligence for microcontroller programming

AbstractMicrocontrollers are widely used in agriculture, yet most undergraduate agriculture students do not have the programming skills necessary to make use of these devices in their academic programs or careers. However, generative artificial intelligence (AI) chatbots, such as ChatGPT, have the ability to write complex microcontroller programs when properly queried. The study was conducted to determine the effects of undergraduate agriculture students’ (n = 22) use of ChatGPT to write a microcontroller program on their programming task performance, self‐efficacy, and attitudes toward generative AI. Nine of 11 (81.8%) student pairs were successful in the ChatGPT‐assisted programming activity, requiring between one (33.3%) and six (11.1%) queries to develop their programs. The two unsuccessful pairs used either one or two queries and produced somewhat functional programs that did not fully operate as specified. Pre‐ and posttest surveys indicated significant (p < 0.001) increases in self‐efficacy for writing microcontroller programs, for using ChatGPT to write microcontroller programs, and attitudes toward generative AI. This research confirmed that undergraduate agriculture students can successfully use generative AI chatbots to write microcontroller programs and that successful task completion increases student self‐efficacy. Further research is needed to determine best practices for using generative AI in teaching and learning microcontroller programming.

Application of large language models in engineering education: A case study of system modeling and simulation courses

In modern engineering education, the application of digital technologies has significantly improved teaching effectiveness and student learning experiences. This study explores the innovative use of large language models (LLMs) in system modeling and simulation courses. Specifically, LLMs were applied to assist in MATLAB programming tasks, allowing students to learn MATLAB commands and programming techniques more conveniently. Additionally, interactions with LLMs guided students in acquiring cross-disciplinary knowledge related to modeling and simulation. In the context of system modeling and control problems, LLMs were utilized to aid in mathematical logic analysis and reasoning by providing potential solutions. These measures have demonstrated that students’ understanding and mastery of complex concepts were improved, and their interest and initiative in learning were stimulated. This paper summarizes the experiences of integrating LLMs in teaching, discusses their potential advantages and challenges in engineering education, and highlights the importance of incorporating digital technologies, particularly large language models, to support educational innovation.

Integrating Generative AI in Data Science Programming: Group Differences in Hint Requests

Generative AI applications have increasingly gained visibility in recent educational literature. Yet less is known about how access to generative tools, such as ChatGPT, influences help-seeking during complex problem-solving. In this paper, we aim to advance the understanding of learners’ use of a support strategy (hints) when solving data science programming tasks in an online AI-enabled learning environment. The study compared two conditions: students solving problems in DaTu with AI assistance (N=45) and those without AI assistance (N=44). Findings reveal no difference in hint-seeking behavior between the two groups, suggesting that the integration of AI assistance has minimal impact on how individuals seek help. The findings also suggest that the availability of AI assistance does not necessarily reduce learners’ reliance on support strategies (such as hints). The current study advances data science education and research by exploring the influence of AI assistance during complex data science problem-solving. We discuss implications and identify paths for future research.

Economic model predictive control for building HVAC system: A comparative analysis of model-based and data-driven approaches using the BOPTEST Framework

Model Predictive Control (MPC) has demonstrated its capability to significantly enhance energy efficiency while ensuring indoor comfort. This study focuses on the development and fair comparison of two Economic Model Predictive Control(EMPC) frameworks with the goal of optimizing demand response for utility cost reduction and ensuring thermal comfort in residential buildings. The first framework, known as RC-MPC, utilizes a reduced-order thermal Resistance–Capacitance network model to represent the thermal dynamics of the building. It formulates the optimization problem as a Linear Programming (LP) task, which is solved using CPLEX. The second framework, known as ANN-MPC, employs an Artificial Neural Network model capture the thermal behavior of the building, It addresses a non-convex optimization problem through projected stochastic gradient descent. To evaluate the performance of these frameworks, experiments are conducted using a single-zone dwelling within the BOPTEST framework. The results demonstrate that both approaches effectively reduce operational costs and enhance comfort, achieving comparable performance and outperforming the baseline Rule-Based Control (RBC) method. The discomfort reductions range from 30% to 95%, and energy expense savings range from 17% to 34%. Furthermore, the study examines the impact of various MPC hyperparameters. The findings highlight that a proper control horizon and formulating the objective function accurately are crucial for achieving optimal control performance. Additionally, tightening the comfort range is found to be important in managing the comfort violations, and the choice of solver can influence both the control performance and computational efficiency.

Returning value from the All of Us research program to PhD-level nursing students using ChatGPT as programming support: results from a mixed-methods experimental feasibility study.

We aimed to evaluate the feasibility of using ChatGPT as programming support for nursing PhD students conducting analyses using the All of Us Researcher Workbench. 9 students in a PhD-level nursing course were prospectively randomized into 2 groups who used ChatGPT for programming support on alternating assignments in the workbench. Students reported completion time, confidence, and qualitative reflections on barriers, resources used, and the learning process. The median completion time was shorter for novices and certain assignments using ChatGPT. In qualitative reflections, students reported ChatGPT helped generate and troubleshoot code and facilitated learning but was occasionally inaccurate. ChatGPT provided cognitive scaffolding that enabled students to move toward complex programming tasks using the All of Us Researcher Workbench but should be used in combination with other resources. Our findings support the feasibility of using ChatGPT to help PhD nursing students use the All of Us Researcher Workbench to pursue novel research directions.

A survey analysis of the adoption of large language models among pathologists.

We sought to investigate the adoption and perception of large language model (LLM) applications among pathologists. A cross-sectional survey was conducted, gathering data from pathologists on their usage and views concerning LLM tools. The survey, distributed globally through various digital platforms, included quantitative and qualitative questions. Patterns in the respondents' adoption and perspectives on these artificial intelligence tools were analyzed. Of 215 respondents, 100 (46.5%) reported using LLMs, particularly ChatGPT (OpenAI), for professional purposes, predominantly for information retrieval, proofreading, academic writing, and drafting pathology reports, highlighting a significant time-saving benefit. Academic pathologists demonstrated a better level of understanding of LLMs than their peers. Although chatbots sometimes provided incorrect general domain information, they were considered moderately proficient concerning pathology-specific knowledge. The technology was mainly used for drafting educational materials and programming tasks. The most sought-after feature in LLMs was their image analysis capabilities. Participants expressed concerns about information accuracy, privacy, and the need for regulatory approval. Large language model applications are gaining notable acceptance among pathologists, with nearly half of respondents indicating adoption less than a year after the tools' introduction to the market. They see the benefits but are also worried about these tools' reliability, ethical implications, and security.

Integrated reading-into-writing: Developing a task blueprint

Integrated reading-into-writing assessment has become increasingly common in second language English for Academic Purposes assessment. It is viewed as a means of potentially increasing the authenticity and therefore validity of assessment that aims to measure a candidate’s readiness for the demands of academic writing. While there is a body of research seeking to identify the distinct sub-skills involved in an integrated reading to write task, there is much more limited work addressing the design of rubrics for integrated tasks for specific contexts. This paper describes the development of an integrated reading-into-writing task and rubric for students at B2+/C1 CEFR level in a PreMasters Direct Entry Program (DEP) at a university language centre in Australia. The project aimed to develop a blueprint for an integrated reading-into-writing task and rubric that captures the core constructs of synthesis writing and can be used as a basis for assessment writers to adapt to suit the parameters of other specific integrated reading into writing tasks in other programs or contexts.

Multilingual Students’ Solution Strategies in Solving Linear Programming Problems: A Case of National Curriculum Vocational Level 3 Mathematics Students

This paper presents an exploratory case study that investigates multilingual students' solution strategies for linear programming problems at a technical vocational education and training (TVET) college in Gauteng. South Africa, like many developing nations, uses English as the language of instruction in technical vocational education and training in TVET colleges, despite the fact that many students lack the essential fluency to interact effectively with the curriculum. Where TVET lecturers and students speak the same home language, there is a disconnect between language policy and codeswitching, which is a frequent approach for achieving pedagogical goals. However, lecturers' training in multilingual realms in the classroom has frequently been framed in terms of linguistic issues, with a limited understanding of code-switching. The ability to combine two languages in the classroom in a systematic manner to promote learning has not been widely appreciated. Furthermore, codeswitching discourses are frequently veiled, with lecturers bringing the vernacular into the classroom. The authors of this study advocate for the discontinuity framework, which claims that in order to learn mathematics, pupils must first acquire the essential language. A purposive sample of 18 National Certificate (Vocational) Level 3 bilingual students was given the linear programming challenge. Data were gathered through a test and semi-structured interviews. The inductive content analysis, which was used for collection and analysis, revealed that students were unable to complete the linear programming task successfully because to a lack of relevant linear programming vocabulary (register) caused by language barriers. We thus recommend that lecturers use the mathematics register in their discussions with students about mathematics to demonstrate the complex and precise ways of expressing mathematical ideas; for example, lectures may re-voice their students' language representations so that these expressions more closely approximate the precision of the mathematics register.

РЕАЛІЗАЦІЯ ДИФЕРЕНЦІЙОВАНОГО ПІДХОДУ ПРИ НАВЧАННІПРОГРАМУВАННЮ МОВОЮ PYTHON ЗДОБУВАЧІВЗАГАЛЬНОЇ СЕРЕДНЬОЇ ОСВІТИ

Students choose their future path starting from school. The learner can determine their abilities when they genuinely understand their potential and prospects in a specific subject or direction. A differentiated approach helps explore the limits of knowledge and encourages its expansion. The paper discusses the implementation of a differentiated approach in teaching computer science, particularly in programming. Tasks are proposed to students, emphasizing the method of solving. Each learner chooses the method that suits them best. Evaluation criteria are provided in advance, helping students identify the aspects of the task to focus on. An example task demonstrating the implementation of a differentiated approach involves calculating the recommended body weight for a woman and a man using the Broca's Formula. Depending on the complexity chosen by the student, this task can be solved using a linear structure, applying branching, or developing a widget. Widgets allow students to demonstrate the use of libraries and functions in the Python programming language. Various solutions in Python and recommendations for the placement of such tasks in the study of a specific topic are also provided. During the defense, all students become acquainted with different task implementations, allowing each student to understand various execution methods and choose a more challenging one next time. The presented example can be extended to other topics, in the study of other programming languages or programming environments. Prospects for further research include developing a system of multi-level programming tasks for the purpose of implementing differentiated learning

SCC-GPT: Source Code Classification Based on Generative Pre-Trained Transformers

Developers often rely on online resources, such as Stack Overflow (SO), to seek assistance for programming tasks. To facilitate effective search and resource discovery, manual tagging of questions and posts with the appropriate programming language is essential. However, accurate tagging is not consistently achieved, leading to the need for the automated classification of code snippets into the correct programming language as a tag. In this study, we introduce a novel approach to automated classification of code snippets from Stack Overflow (SO) posts into programming languages using generative pre-trained transformers (GPT). Our method, which does not require additional training on labeled data or dependency on pre-existing labels, classifies 224,107 code snippets into 19 programming languages. We employ the text-davinci-003 model of ChatGPT-3.5 and postprocess its responses to accurately identify the programming language. Our empirical evaluation demonstrates that our GPT-based model (SCC-GPT) significantly outperforms existing methods, achieving a median F1-score improvement that ranges from +6% to +31%. These findings underscore the effectiveness of SCC-GPT in enhancing code snippet classification, offering a cost-effective and efficient solution for developers who rely on SO for programming assistance.

The influence of multiple cognitive workload levels of an exergame on dorsal attention network connectivity at the source level

This study investigates how adding a cognitive task on a balance board (exergame) affects connectivity in the dorsal attention network (DAN) during an exergame task. Healthy young adults performed a soccer ball-moving task by tilting a balance board with their feet while their brain activity was measured using electroencephalography (EEG). In this exergame, the speed of obstacles in front of the goal manipulated the cognitive workload. Higher speed means a higher cognitive workload. The study found significant changes in functional connectivity within DAN regions, specifically in the alpha band. During the shift from easy to medium cognitive task, we observed a significant increase in connectivity (p= 0.0436) between the right inferior temporal (ITG R) and the Left middle temporal (MTG L). During the transition from easy to hard cognitive tasks, strengthened interactions (p= 0.0324) between inferior temporal (ITG) and parsopercularis (pOPPER) were found. This suggests that the proposed balanceboard-based exergame enhances the functionality of specific brain regions, such as ITG and MTG regions, and improves connectivity in the frontal cortex. We also found a correlation between brain activity and performance data, highlighting that increased cognitive workload resulted in decreased performance and heightened frontal alpha activity. These findings align with research suggesting that adding cognitive games to physical activity-based tasks in rehabilitation programs can boost brain activity, resulting in improved decision-making and visual processing skills. This information can help clinicians tailor rehabilitation methods that target specific brain regions.

Pengujian Usabilitas Pada Penggunaan Platform Scratch

This research tests the usability of the platform on new students of Informatics Engineering. Usability of is measured through three main variables: effectiveness, efficiency, and user satisfaction. Data was collected through pre-test and post-test involving 20 respondents using questionnaires and practical tasks. 20 respondents using questionnaires and practical tasks. The results of the study showed that the effectiveness and efficiency of the Scratch platform increased significantly from the pre-test to the post-test, reaching a level of effectiveness and efficiency, showing that the time required to complete programming tasks is reduced once users are familiar with the platform. User satisfaction also increased, with the majority of respondents feeling more comfortable and interested in using Scratch after testing. This research concludes that the Scratch platform is effective and efficient as a programming learning tool for new Informatics Engineering students.

Digital Twins-Based Cognitive Apprenticeship Model in Smart Agriculture

Applying digital technologies in education has been proven to have several positive impacts on teaching and learning. Digital twins have the potential to revolutionize education by offering immersive and interactive learning experiences. They can simulate virtual environments, enabling students to conduct experiments, practice techniques, and observe outcomes in a safe and cost-effective setting, particularly when access to physical laboratories is limited or expensive. This work proposes a learning management model that applies cognitive apprenticeship learning theory as a framework for learning digital twins to enhance the programming of a smart agriculture system and computational thinking skills. The model consists of three parts: input, learning process, and output with feedback. The input phase includes setting learning objectives, analyzing learners, determining learning content, planning learning activities, and utilizing a digital twin virtual laboratory space for experimenting in the development of a smart agriculture system. The learning process involves engaging students in programming and problem-solving tasks by offering expert guidance and support in computationally solving problems. The achievement of learning outcomes and practices is included in the output and feedback sections. After evaluation, five experts in this subject agreed that the learning management model was at the most appropriate level.

Decomposition-based Synthesis for Applying Divide-and-Conquer-like Algorithmic Paradigms

Algorithmic paradigms such as divide-and-conquer (D&C) are proposed to guide developers in designing efficient algorithms, but it can still be difficult to apply algorithmic paradigms to practical tasks. To ease the usage of paradigms, many research efforts have been devoted to the automatic application of algorithmic paradigms. However, most existing approaches to this problem rely on syntax-based program transformations and thus put significant restrictions on the original program. In this article, we study the automatic application of D&C and several similar paradigms, denoted as D&C-like algorithmic paradigms, and aim to remove the restrictions from syntax-based transformations. To achieve this goal, we propose an efficient synthesizer, named AutoLifter , which does not depend on syntax-based transformations. Specifically, the main challenge of applying algorithmic paradigms is from the large scale of the synthesized programs, and AutoLifter addresses this challenge by applying two novel decomposition methods that do not depend on the syntax of the input program, component elimination and variable elimination , to soundly divide the whole problem into simpler subtasks, each synthesizing a sub-program of the final program and being tractable with existing synthesizers. We evaluate AutoLifter on 96 programming tasks related to six different algorithmic paradigms. AutoLifter solves 82/96 tasks with an average time cost of 20.17 s, significantly outperforming existing approaches.

Humanoid Robots as Learning Assistants? Useability Perspectives of Grade 6 Students

AbstractDespite the increasing number of robots in educational settings, they are relatively new, with ambiguous research findings regarding a robot’s influence on students’ learning, psychosocial aspects, and gender differences. In this study, 17 Grade 6 students in a Norwegian school participated in a human–robot interaction (HRI) session with mathematics and programming tasks followed by focus group interviews to explore their experiences. Thematic analysis resulted in four main themes: (1) Technical aspects, (2) Robots as educational assistants, (3) Robot features, and (4) Attitudes regarding robots. Overall, the students found the robot entertaining, providing mastery experiences in mathematics and programming. Some technical difficulties were experienced, and the robot’s lack of facial expressions, emotions, and tailored feedback created interaction challenges. Gender differences regarding interest in people versus things were found. The implementation of robots in education shows potential, however, more research is needed, especially longitudinal exploring the impact of HRI in education.

Refining ChatGPT-Generated Code: Characterizing and Mitigating Code Quality Issues

Since its introduction in November 2022, ChatGPT has rapidly gained popularity due to its remarkable ability in language understanding and human-like responses. ChatGPT, based on GPT-3.5 architecture, has shown great promise for revolutionizing various research fields, including code generation. However, the reliability and quality of code generated by ChatGPT remain unexplored, raising concerns about potential risks associated with the widespread use of ChatGPT-driven code generation. In this article, we systematically study the quality of 4,066 ChatGPT-generated programs of code implemented in two popular programming languages, i.e., Java and Python, for 2,033 programming tasks. The goal of this work is threefold. First, we analyze the correctness of ChatGPT on code generation tasks and uncover the factors that influence its effectiveness, including task difficulty, programming language, time that tasks are introduced, and program size. Second, we identify and characterize potential issues with the quality of ChatGPT-generated code. Last, we provide insights into how these issues can be mitigated. Experiments highlight that out of 4,066 programs generated by ChatGPT, 2,756 programs are deemed correct, 1,082 programs provide wrong outputs, and 177 programs contain compilation or runtime errors. Additionally, we further analyze other characteristics of the generated code through static analysis tools, such as code style and maintainability, and find that 1,930 ChatGPT-generated code snippets suffer from maintainability issues. Subsequently, we investigate ChatGPT’s self-repairing ability and its interaction with static analysis tools to fix the errors uncovered in the previous step. Experiments suggest that ChatGPT can partially address these challenges, improving code quality by more than 20%, but there are still limitations and opportunities for improvement. Overall, our study provides valuable insights into the current limitations of ChatGPT and offers a roadmap for future research and development efforts to enhance the code generation capabilities of artificial intelligence models such as ChatGPT.

Deep reinforcement learning based latency-energy minimization in smart healthcare network

Significant breakthroughs in the Internet of Things (IoT) and 5G technologies have driven several smart healthcare activities, leading to a flood of computationally intensive applications in smart healthcare networks. Mobile Edge Computing (MEC) is considered as an efficient solution to provide powerful computing capabilities to latency or energy sensitive nodes. The low-latency and high-reliability requirements of healthcare application services can be met through optimal offloading and resource allocation for the computational tasks of the nodes. In this study, we established a system model consisting of two types of nodes by considering nondivisible and trade-off computational tasks between latency and energy consumption. To minimize processing cost of the system tasks, a Mixed-Integer Nonlinear Programming (MINLP) task offloading problem is proposed. Furthermore, this problem is decomposed into task offloading decisions and resource allocation problems. The resource allocation problem is solved using traditional optimization algorithms, and the offloading decision problem is solved using a deep reinforcement learning algorithm. We propose an Online Offloading based on the Deep Reinforcement Learning (OO-DRL) algorithm with parallel deep neural networks and a weight-sensitive experience replay mechanism. Simulation results show that, compared with several existing methods, our proposed algorithm can perform real-time task offloading in a smart healthcare network in dynamically varying environments and reduce the system task processing cost.

ACcoding: A graph-based dataset for online judge programming

A well-designed educational programming dataset is a valuable asset for students and educators. Such a dataset enables students to improve their programming performances continuously, provides researchers with significant data sources to identify students’ learning behaviours and enhance the quality of programming education. Several existing datasets for programming education are either limited by a small number of participating students or a short span of learning records, bringing great challenges to investigate students’ learning patterns in programming. We present a graph-based large-scale dataset specialized in programming learning on Online Judge (OJ) platform. The dataset, named ACcoding, was built by a university teaching group. As of the submission date of the initial manuscript of this paper (May 6, 2022), the dataset contains 4,046,652 task-solving records submitted by 27,444 students on 4,559 programming tasks over a span of 6 years. The large size of the dataset, combined with rich functional features, empowers educators to trace students’ programming progress and choose appropriate programming tasks for specific training purposes. We also presents examples of applications used by the dataset.