This study evaluates the effectiveness of four leading large language models (LLMs), GPT-3, GPT-4, GPT-4o, and BERT, in generating quiz questions for Java and Python programming courses. We aim to recognize how LLMs c...
详细信息
Purpose: Over the last year, the ascent of Generative AI (GenAI) has raised concerns about its impact on core skill development, such as problem-solving and algorithmic thinking, in computerscience students. With the...
详细信息
Purpose: Over the last year, the ascent of Generative AI (GenAI) has raised concerns about its impact on core skill development, such as problem-solving and algorithmic thinking, in computerscience students. With the proliferation of these tools, educators must evaluate their role in academia, as neglecting this might culminate in a phenomenon we term the “Junior-Year Wall,” where students struggle in advanced courses due to prior over-dependence on GenAI. Our research seeks to answer the question: “How can educators guide students’ interactions with GenAI to preserve core skill development during their foundational academic years?” Methods: We introduce “AI-Lab,” a pedagogical framework for guiding students in effectively leveraging GenAI within core collegiate programming courses. This framework accentuates GenAI’s benefits and potential as a pedagogical instrument. Specifically, AI-Lab presents opportunities to use GenAI for tailored support, such as topic introductions, detailed examples, corner case identification, rephrased explanations, and debugging assistance. Through identifying and rectifying GenAI’s errors, students enrich their learning process. Additionally, AI-Lab offers strategies for formulating prompts to elicit high-quality GenAI responses and provides mechanisms for educators to explore students’ perceptions of GenAI’s role in their learning experience. Results: Preliminary anonymous surveys show that at least 54.5% of our students use GenAI for homework. Notably, the framework highlights the risks of GenAI over-dependence as well as introducing its context-specific benefits, aiming to motivate students intrinsically towards balanced usage. Conclusions: The AI-Lab framework underscores the importance of guiding students’ interactions with GenAI to maintain core skill development in computerscience. By fostering an environment where GenAI is used as a pedagogical tool, educators can mitigate the risks associated with over-dependence on GenAI. Th
Currently, there are a plethora of solutions developed to help students learn the basics of programming. However, there is a relative paucity of solutions that cater to problems students face when learning programming...
详细信息
Currently, there are a plethora of solutions developed to help students learn the basics of programming. However, there is a relative paucity of solutions that cater to problems students face when learning programming that is mainly caused by the abstract nature of programming, misconceptions of programming concepts, and lack of motivation. Hence, in this study, a framework to address the abstract nature of programming and common programming misconceptions is developed. The framework consists of three modules that correspond to each issue, powered by a simulation engine. The first module is developed to address the abstract nature of programming by representing programming concepts with concrete objects in the virtual environment. The second module employs simulation techniques such as interactions and player perspectives to address common programming misconceptions. Lastly, the third module employs elements in the virtual environment to engage students when learning through the system. To evaluate the system, 60 participants were randomly divided into the control group (N = 30) and the experimental group (N = 30). Participants in the control group were taught using a video lecture while participants in the experimental group were taught using the developed VR intervention. Evaluation results gathered quantitatively indicated that the VR intervention was able to significantly increase programming concepts comprehension and address programming misconceptions. Participants also rated the developed VR intervention to be significantly more engaging than the video lecture.
The importance and growth of the Internet of Things (IoT) in computer networks and applications have been increasing. Additionally, many of these applications generate large volumes of data, which are critical and req...
详细信息
Pruning is a major research field in neural networks, enhancing their efficiency and generalization. The field of pruning approaches in genetic programming (GP) is continually evolving, with researchers actively explo...
详细信息
This research explores the links between self-regulation behaviors and indicators of learning performance. A data mining approach coupled with appropriate qualitative measures is proposed to extract behavioral sequenc...
详细信息
ISBN:
(纸本)9783031630309;9783031630316
This research explores the links between self-regulation behaviors and indicators of learning performance. A data mining approach coupled with appropriate qualitative measures is proposed to extract behavioral sequences that are representative of learning success. Applied on an online programming platform, obtained results allowed to highlight important self-regulation behaviors during the planning and engagement phases. It e.g. appears that successful self-regulated learners are those who analyze their tasks before working on them. This work brings methodological contributions in the field of self-regulation learning measurement and is a first step towards the design of intelligent tutoring systems.
We study linear bilevel programming problems whose lower-level objective is given by a random cost vector with known distribution. We consider the case where this distribution is nonatomic, allowing to pose the proble...
详细信息
ISBN:
(纸本)9783031327254;9783031327261
We study linear bilevel programming problems whose lower-level objective is given by a random cost vector with known distribution. We consider the case where this distribution is nonatomic, allowing to pose the problem of the leader using vertex-supported beliefs in the sense of [29]. We prove that, under suitable assumptions, this formulation turns out to be piecewise affine over the so-called chamber complex of the feasible set of the high point relaxation. We propose two algorithmic approaches to solve general problems enjoying this last property. The first one is based on enumerating the vertices of the chamber complex. The second one is a Monte-Carlo approximation scheme based on the fact that randomly drawn points of the domain lie, with probability 1, in the interior of full-dimensional chambers, where the problem (restricted to this chamber) can be reduced to a linear program.
ChatGPT is a conversational AI platform that can produce code to solve problems when provided with a natural language prompt. Prior work on similar AI models has shown that they perform well on typical intro-level Com...
详细信息
ISBN:
(纸本)9798400704239
ChatGPT is a conversational AI platform that can produce code to solve problems when provided with a natural language prompt. Prior work on similar AI models has shown that they perform well on typical intro-level computerscience problems. However, little is known about the performance of such tools on Data science (DS) problems. In this work, we assess the performance of ChatGPT on assignments from three DS courses with varying difficulty levels. First, we apply the raw assignment prompts provided to the students and find that ChatGPT performs well on assignments with dataset(s) descriptions and progressive question prompts, which divide the programming requirements into sub-problems. Then, we perform prompt engineering on the assignments for which ChatGPT had low performance. We find that the following prompt engineering techniques significantly increased ChatGPT's performance: breaking down abstract questions into steps, breaking down steps into multiple prompts, providing descriptions of the dataset(s), including algorithmic details, adding specific instructions to entice specific actions, and removing extraneous information. Finally, we discuss how our findings suggest potential changes to curriculum design of DS courses.
When learning programming, it is important to improve one's own code. In this study, we have proposed an evaluation function that uses robot programming and a ranking system based on evaluations, and have develope...
详细信息
ISBN:
(纸本)9783031351280;9783031351297
When learning programming, it is important to improve one's own code. In this study, we have proposed an evaluation function that uses robot programming and a ranking system based on evaluations, and have developed and evaluated a code-sharing platform that allows users to learn only code with similar ranks. Although the evaluation confirmed a certain learning effect of the developed system in an experimental environment with a small number of participants, it is desirable to increase the number of participants in an environment where the code is shared. Therefore in this paper, we conducted a class for second-year university students and verified whether the same learning effects could be obtained in a large-group environment. The evaluation results suggest that the learning effect and the motivation to learn are enhanced in the same way as in a small-group environment.
Early skin cancer diagnosis saves lives as the disease can be successfully treated through complete excision. computer-aided diagnosis methods are developed using artificial intelligence techniques to help earlier det...
详细信息
Early skin cancer diagnosis saves lives as the disease can be successfully treated through complete excision. computer-aided diagnosis methods are developed using artificial intelligence techniques to help earlier detection and identify hidden causes leading to cancers in skin lesion images. In skin cancer image classification problems, an ensemble of classifiers has demonstrated better classification ability than a single classification algorithm. Traditionally, training an ensemble uses the complete set of original features, where some of these features can be redundant or irrelevant and hence, may not provide useful information in generating good models for ensemble classification. Moreover, newly created features may help improve classification performance. To address this issue, the existing methods have used feature construction for building an ensemble classifier, which usually creates a fixed number of features that may fit the training data too well, resulting in poor test performance. This study develops a novel classification approach that combines ensemble learning, feature selection, and feature construction utilizing genetic programming (GP) to handle the above limitations. The proposed method automatically evolves variable-length feature vectors consisting of GP-selected and GP-constructed features suitable for training an ensemble classifier. This study evaluates the effectiveness of the proposed method on two benchmark real-world skin image datasets that include dermoscopy and standard camera images. The experimental results reveal that the proposed algorithm significantly outperforms four state-of-the-art convolutional neural network methods, the existing GP approaches, and 11 commonly used machine learning methods. Furthermore, this study also includes interpreting evolved individuals that highlight important skin cancer characteristics playing a vital role in discriminating images of different cancer classes. This study shows that high classifica
暂无评论