Arto Hellas

Charles Koutcheme, Nicola Dainese, Arto Hellas et al.

The emergence of large language models (LLMs) has transformed research and practice across a wide range of domains. Within the computing education research (CER) domain, LLMs have garnered significant attention, particularly in the context of learning programming. Much of the work on LLMs in CER, however, has focused on applying and evaluating proprietary models. In this article, we evaluate the efficiency of open-source LLMs in generating high-quality feedback for programming assignments and judging the quality of programming feedback, contrasting the results with proprietary models. Our evaluations on a dataset of students' submissions to introductory Python programming exercises suggest that state-of-the-art open-source LLMs are nearly on par with proprietary models in both generating and assessing programming feedback. Additionally, we demonstrate the efficiency of smaller LLMs in these tasks and highlight the wide range of LLMs accessible, even for free, to educators and practitioners.

Sami Sarsa, Paul Denny, Arto Hellas et al.

This article explores the natural language generation capabilities of large language models with application to the production of two types of learning resources common in programming courses. Using OpenAI Codex as the large language model, we create programming exercises (including sample solutions and test cases) and code explanations, assessing these qualitatively and quantitatively. Our results suggest that the majority of the automatically generated content is both novel and sensible, and in some cases ready to use as is. When creating exercises we find that it is remarkably easy to influence both the programming concepts and the contextual themes they contain, simply by supplying keywords as input to the model. Our analysis suggests that there is significant value in massive generative machine learning models as a tool for instructors, although there remains a need for some oversight to ensure the quality of the generated content before it is delivered to students. We further discuss the implications of OpenAI Codex and similar tools for introductory programming education and highlight future research streams that have the potential to improve the quality of the educational experience for both teachers and students alike.

Paul Denny, James Prather, Brett A. Becker et al.

The computing education community has a rich history of pedagogical innovation designed to support students in introductory courses, and to support teachers in facilitating student learning. Very recent advances in artificial intelligence have resulted in code generation models that can produce source code from natural language problem descriptions -- with impressive accuracy in many cases. The wide availability of these models and their ease of use has raised concerns about potential impacts on many aspects of society, including the future of computing education. In this paper, we discuss the challenges and opportunities such models present to computing educators, with a focus on introductory programming classrooms. We summarize the results of two recent articles, the first evaluating the performance of code generation models on typical introductory-level programming problems, and the second exploring the quality and novelty of learning resources generated by these models. We consider likely impacts of such models upon pedagogical practice in the context of the most recent advances at the time of writing.

Juho Leinonen, Arto Hellas, Sami Sarsa et al.

A key part of learning to program is learning to understand programming error messages. They can be hard to interpret and identifying the cause of errors can be time-consuming. One factor in this challenge is that the messages are typically intended for an audience that already knows how to program, or even for programming environments that then use the information to highlight areas in code. Researchers have been working on making these errors more novice friendly since the 1960s, however progress has been slow. The present work contributes to this stream of research by using large language models to enhance programming error messages with explanations of the errors and suggestions on how to fix the error. Large language models can be used to create useful and novice-friendly enhancements to programming error messages that sometimes surpass the original programming error messages in interpretability and actionability. These results provide further evidence of the benefits of large language models for computing educators, highlighting their use in areas known to be challenging for students. We further discuss the benefits and downsides of large language models and highlight future streams of research for enhancing programming error messages.

Juho Leinonen, Paul Denny, Stephen MacNeil et al.

Reasoning about code and explaining its purpose are fundamental skills for computer scientists. There has been extensive research in the field of computing education on the relationship between a student's ability to explain code and other skills such as writing and tracing code. In particular, the ability to describe at a high-level of abstraction how code will behave over all possible inputs correlates strongly with code writing skills. However, developing the expertise to comprehend and explain code accurately and succinctly is a challenge for many students. Existing pedagogical approaches that scaffold the ability to explain code, such as producing exemplar code explanations on demand, do not currently scale well to large classrooms. The recent emergence of powerful large language models (LLMs) may offer a solution. In this paper, we explore the potential of LLMs in generating explanations that can serve as examples to scaffold students' ability to understand and explain code. To evaluate LLM-created explanations, we compare them with explanations created by students in a large course ($n \approx 1000$) with respect to accuracy, understandability and length. We find that LLM-created explanations, which can be produced automatically on demand, are rated as being significantly easier to understand and more accurate summaries of code than student-created explanations. We discuss the significance of this finding, and suggest how such models can be incorporated into introductory programming education.

Arto Hellas, Juho Leinonen, Sami Sarsa et al.

Background and Context: Over the past year, large language models (LLMs) have taken the world by storm. In computing education, like in other walks of life, many opportunities and threats have emerged as a consequence. Objectives: In this article, we explore such opportunities and threats in a specific area: responding to student programmers' help requests. More specifically, we assess how good LLMs are at identifying issues in problematic code that students request help on. Method: We collected a sample of help requests and code from an online programming course. We then prompted two different LLMs (OpenAI Codex and GPT-3.5) to identify and explain the issues in the students' code and assessed the LLM-generated answers both quantitatively and qualitatively. Findings: GPT-3.5 outperforms Codex in most respects. Both LLMs frequently find at least one actual issue in each student program (GPT-3.5 in 90% of the cases). Neither LLM excels at finding all the issues (GPT-3.5 finding them 57% of the time). False positives are common (40% chance for GPT-3.5). The advice that the LLMs provide on the issues is often sensible. The LLMs perform better on issues involving program logic rather than on output formatting. Model solutions are frequently provided even when the LLM is prompted not to. LLM responses to prompts in a non-English language are only slightly worse than responses to English prompts. Implications: Our results continue to highlight the utility of LLMs in programming education. At the same time, the results highlight the unreliability of LLMs: LLMs make some of the same mistakes that students do, perhaps especially when formatting output as required by automated assessment systems. Our study informs teachers interested in using LLMs as well as future efforts to customize LLMs for the needs of programming education.

Paul Denny, Hassan Khosravi, Arto Hellas et al.

As an increasing number of students move to online learning platforms that deliver personalized learning experiences, there is a great need for the production of high-quality educational content. Large language models (LLMs) appear to offer a promising solution to the rapid creation of learning materials at scale, reducing the burden on instructors. In this study, we investigated the potential for LLMs to produce learning resources in an introductory programming context, by comparing the quality of the resources generated by an LLM with those created by students as part of a learnersourcing activity. Using a blind evaluation, students rated the correctness and helpfulness of resources generated by AI and their peers, after both were initially provided with identical exemplars. Our results show that the quality of AI-generated resources, as perceived by students, is equivalent to the quality of resources generated by their peers. This suggests that AI-generated resources may serve as viable supplementary material in certain contexts. Resources generated by LLMs tend to closely mirror the given exemplars, whereas student-generated resources exhibit greater variety in terms of content length and specific syntax features used. The study highlights the need for further research exploring different types of learning resources and a broader range of subject areas, and understanding the long-term impact of AI-generated resources on learning outcomes.

Rose Niousha, Samantha Boatright Smith, Bita Akram et al.

Current Artificial Intelligence (AI)-based tutoring systems (AI tutors) are primarily evaluated based on the pedagogical quality of their feedback messages. While important, pedagogy alone is insufficient because it ignores a critical question: what do students actually do with the feedback they receive? We argue that AI tutor evaluation should be extended with a behavioral dimension grounded in student interaction data, which complements pedagogical assessment. We propose an evaluation framework and apply it to 10,235 code submissions with corresponding AI tutor feedback from an introductory undergraduate programming course to measure whether students act on tutor feedback and whether those actions are applied correctly. Using this framework to compare two deployed AI tutors across different semesters in a large-scale introductory computer science course reveals substantial differences in student engagement patterns that are not captured by pedagogy-only evaluation. Moreover, these engagement-based behavioral signals are more strongly associated with student perception of helpful feedback than pedagogical quality alone, providing a more complete and actionable picture of AI tutor performance.

Stephen MacNeil, Paul Denny, Andrew Tran et al.

Identifying and resolving logic errors can be one of the most frustrating challenges for novices programmers. Unlike syntax errors, for which a compiler or interpreter can issue a message, logic errors can be subtle. In certain conditions, buggy code may even exhibit correct behavior -- in other cases, the issue might be about how a problem statement has been interpreted. Such errors can be hard to spot when reading the code, and they can also at times be missed by automated tests. There is great educational potential in automatically detecting logic errors, especially when paired with suitable feedback for novices. Large language models (LLMs) have recently demonstrated surprising performance for a range of computing tasks, including generating and explaining code. These capabilities are closely linked to code syntax, which aligns with the next token prediction behavior of LLMs. On the other hand, logic errors relate to the runtime performance of code and thus may not be as well suited to analysis by LLMs. To explore this, we investigate the performance of two popular LLMs, GPT-3 and GPT-4, for detecting and providing a novice-friendly explanation of logic errors. We compare LLM performance with a large cohort of introductory computing students $(n=964)$ solving the same error detection task. Through a mixed-methods analysis of student and model responses, we observe significant improvement in logic error identification between the previous and current generation of LLMs, and find that both LLM generations significantly outperform students. We outline how such models could be integrated into computing education tools, and discuss their potential for supporting students when learning programming.

Charles Koutcheme, Arto Hellas, Juho Leinonen

Artificial models that simulate how learners act and respond within educational systems are a promising tool for evaluating tutoring strategies and feedback mechanisms at scale. However, many existing approaches in programming education rely on prompting large, proprietary language models, raising concerns around privacy, cost, and dependence. In this work, we propose a method for training open-weight artificial programming learners using authentic student process data. Our approach serializes temporal log traces into a conversational format, representing each student's problem-solving process as a dialogue between the learner and their automated assessment system. Student code submissions and environment feedback, such as test outcomes, grades, and error traces, form alternating conversational turns, enabling models to learn from the iterative debugging process. We additionally introduce a training pipeline combining supervised fine-tuning with preference optimization to align models with authentic student debugging behavior. We evaluate our framework by training Qwen models at 4B and 8B scales on a large-scale dataset of real student submissions to Python programming assignments. Our results show that incorporating environment feedback strengthens the models' ability to replicate student debugging behavior, improving over both prior code-only approaches and prompted large language models baselines in functional alignment and code similarity. We release our code to support reproducibility.

Evanfiya Logacheva, Arto Hellas, Tsvetomila Mihaylova et al.

Generative artificial intelligence (AI) has found a widespread use in computing education; at the same time, quality of generated materials raises concerns among educators and students. This study addresses this issue by introducing a novel method for diagram code generation with in-context examples based on the Rhetorical Structure Theory (RST), which aims to improve diagram generation by aligning models' output with user expectations. Our approach is evaluated by computer science educators, who assessed 150 diagrams generated with large language models (LLMs) for logical organization, connectivity, layout aesthetic, and AI hallucination. The assessment dataset is additionally investigated for its utility in automated diagram evaluation. The preliminary results suggest that our method decreases the rate of factual hallucination and improves diagram faithfulness to provided context; however, due to LLMs' stochasticity, the quality of the generated diagrams varies. Additionally, we present an in-depth analysis and discussion on the connection between AI hallucination and the quality of generated diagrams, which reveals that text contexts of higher complexity lead to higher rates of hallucination and LLMs often fail to detect mistakes in their output.

Charles Koutcheme, Nicola Dainese, Sami Sarsa et al.

Large language models (LLMs) have shown great potential for the automatic generation of feedback in a wide range of computing contexts. However, concerns have been voiced around the privacy and ethical implications of sending student work to proprietary models. This has sparked considerable interest in the use of open source LLMs in education, but the quality of the feedback that such open models can produce remains understudied. This is a concern as providing flawed or misleading generated feedback could be detrimental to student learning. Inspired by recent work that has utilised very powerful LLMs, such as GPT-4, to evaluate the outputs produced by less powerful models, we conduct an automated analysis of the quality of the feedback produced by several open source models using a dataset from an introductory programming course. First, we investigate the viability of employing GPT-4 as an automated evaluator by comparing its evaluations with those of a human expert. We observe that GPT-4 demonstrates a bias toward positively rating feedback while exhibiting moderate agreement with human raters, showcasing its potential as a feedback evaluator. Second, we explore the quality of feedback generated by several leading open-source LLMs by using GPT-4 to evaluate the feedback. We find that some models offer competitive performance with popular proprietary LLMs, such as ChatGPT, indicating opportunities for their responsible use in educational settings.

Tsvetomila Mihaylova, Evanfiya Logacheva, Arto Hellas et al.

When programming students encounter errors in their code, compiler messages or static analysis output often provide limited guidance, particularly for novice programmers. Personalized feedback from instructors can be effective but does not scale well. Recent advances in large language models (LLMs) enable automated feedback generation at scale. This study examines whether LLM-generated feedback with different levels of guidance is associated with differences in students' problem-solving behavior. We analyze effects on time to solution and number of attempts, and examine whether these effects differ by programming experience. We design three feedback types and compare them to a baseline in which students receive only compiler error messages. Results from an online programming course show that LLM-generated feedback is associated with faster time to solution compared to the no-feedback baseline, with less guided feedback showing slightly stronger effects. Overall, the findings suggest that feedback structure plays an important role in how students progress toward correct solutions and motivate further work on adaptive feedback designs and longer-term learning outcomes.

Juho Leinonen, Lisa Zhang, Arto Hellas

As generative AI (GenAI) tools become easily accessible, there is promise in using such tools to support instructors. To that end, this paper examines using GenAI to help generate slides from instructor authored course notes, emphasizing instructor and student perceptions. We examine an end-to-end education tool (NotebookLM), two general-purpose LLMs (Claude, M365 Copilot), and two coding assistants (Cursor, Claude Code). We first analyze whether GenAI generated slides are ``good'' via narrative assessment by educators. We choose the best slides to use (with some modification) in a real course setting, and compare the student perception of human vs. AI generated slides. We find that coding assistant tools produce slides that were most accurate, complete, and pedagogically sound. Additionally, students rate GenAI slides to be of similar quality as instructor-created slides, and cannot reliably identify which slides are AI-generated. Additionally, we find a negative correlation between a high quality rating and a high ``AI-generated'' rating, suggesting students associate poor quality with the source of the slides being AI. These findings highlight promising opportunities for integrating GenAI into instructional design workflows and call for further research on how educators can best harness such tools responsibly and effectively.

Mragisha Jain, Tirth Bhatt, Griffin Pitts et al.

Students learning algorithms often need support as they interpret traces, debug reasoning errors, and apply procedures across unfamiliar problem instances. In this paper, we present KITE (Knowledge-Informed Tutoring Engine), a Retrieval-Augmented Generation (RAG)-based intelligent tutoring system designed to serve as a classroom teaching assistant for algorithmic reasoning and problem-solving tasks. KITE uses an intent-aware Socratic response strategy to tailor support to different student needs, responding with targeted hints, guiding questions, and progressive scaffolding intended to strengthen students' algorithmic problem-solving ability. To keep responses aligned with course content, KITE uses a multimodal RAG pipeline that retrieves relevant information from course materials. We evaluate KITE using three forms of assessment: RAGAs-based metrics for response grounding and quality, expert evaluation of pedagogical quality, and a simulated student pipeline in which a weaker language model interacts with KITE across two-turn dialogues and produces revised answers after receiving feedback. Results indicate that KITE produces contextually grounded and pedagogically appropriate responses. Further, using simulated students, KITE's feedback helped the student models produce more accurate follow-up responses on procedural and tracing questions, suggesting that its scaffolding can support algorithmic problem-solving. This work contributes a tutoring architecture and an evaluation approach for assessing retrieval-grounded explanations and scaffolded problem-solving feedback.

Zainab Khan, Ahmed Hussain, Mukesh Thakur et al.

The use of Service-Based Architecture in modern telecommunications has exponentially increased Network Functions (NFs) and Application Programming Interfaces (APIs), creating substantial operational complexities in service discovery and management. We introduce \textit{NEFMind}, a framework leveraging parameter-efficient fine-tuning of open-source Large Language Models (LLMs) to address these challenges. It integrates three core components: synthetic dataset generation from Network Exposure Function (NEF) API specifications, model optimization through Quantized-Low-Rank Adaptation, and performance evaluation via GPT-4 Ref Score and BertScore metrics. Targeting 5G Service-Based Architecture APIs, our approach achieves 85% reduction in communication overhead compared to manual discovery methods. Experimental validation using the open-source Phi-2 model demonstrates exceptional API call identification performance at 98-100% accuracy. The fine-tuned Phi-2 model delivers performance comparable to significantly larger models like GPT-4 while maintaining computational efficiency for telecommunications infrastructure deployment. These findings validate domain-specific, parameter-efficient LLM strategies for managing complex API ecosystems in next-generation telecommunications networks.

Griffin Pitts, Muntasir Hoq, Peter Brusilovsky et al.

Adaptive programming practice often relies on fixed libraries of worked examples and practice problems, which require substantial authoring effort and may not correspond well to the logical errors and partial solutions students produce while writing code. As a result, students may receive learning content that does not directly address the concepts they are working to understand, while instructors must either invest additional effort in expanding content libraries or accept a coarse level of personalization. We present an approach for knowledge-component (KC) guided educational content generation using pattern-based KCs extracted from student code. Given a problem statement and student submissions, our pipeline extracts recurring structural KC patterns from students' code through AST-based analysis and uses them to condition a generative model. In this study, we apply this approach to worked example generation, and compare baseline and KC-conditioned outputs through expert evaluation. Results suggest that KC-conditioned generation improves topical focus and relevance to learners' underlying logical errors, providing evidence that KC-based steering of generative models can support personalized learning at scale.

Charles Koutcheme, Nicola Dainese, Sami Sarsa et al.

The emergence of large language models (LLMs) has sparked enormous interest due to their potential application across a range of educational tasks. For example, recent work in programming education has used LLMs to generate learning resources, improve error messages, and provide feedback on code. However, one factor that limits progress within the field is that much of the research uses bespoke datasets and different evaluation metrics, making direct comparisons between results unreliable. Thus, there is a pressing need for standardization and benchmarks that facilitate the equitable comparison of competing approaches. One task where LLMs show great promise is program repair, which can be used to provide debugging support and next-step hints to students. In this article, we propose a novel educational program repair benchmark. We curate two high-quality publicly available programming datasets, present a unified evaluation procedure introducing a novel evaluation metric rouge@k for approximating the quality of repairs, and evaluate a set of five recent models to establish baseline performance.

Stephen MacNeil, Magdalena Rogalska, Juho Leinonen et al.

Large language models (LLMs) present an exciting opportunity for generating synthetic classroom data. Such data could include code containing a typical distribution of errors, simulated student behaviour to address the cold start problem when developing education tools, and synthetic user data when access to authentic data is restricted due to privacy reasons. In this research paper, we conduct a comparative study examining the distribution of bugs generated by LLMs in contrast to those produced by computing students. Leveraging data from two previous large-scale analyses of student-generated bugs, we investigate whether LLMs can be coaxed to exhibit bug patterns that are similar to authentic student bugs when prompted to inject errors into code. The results suggest that unguided, LLMs do not generate plausible error distributions, and many of the generated errors are unlikely to be generated by real students. However, with guidance including descriptions of common errors and typical frequencies, LLMs can be shepherded to generate realistic distributions of errors in synthetic code.

Leo Leppänen, Lili Aunimo, Arto Hellas et al.

The release of ChatGPT in late 2022 caused a flurry of activity and concern in the academic and educational communities. Some see the tool's ability to generate human-like text that passes at least cursory inspections for factual accuracy ``often enough'' a golden age of information retrieval and computer-assisted learning. Some, on the other hand, worry the tool may lead to unprecedented levels of academic dishonesty and cheating. In this work, we quantify some of the effects of the emergence of Large Language Models (LLMs) on online education by analyzing a multi-year dataset of student essay responses from a free university-level MOOC on AI ethics. Our dataset includes essays submitted both before and after ChatGPT's release. We find that the launch of ChatGPT coincided with significant changes in both the length and style of student essays, mirroring observations in other contexts such as academic publishing. We also observe -- as expected based on related public discourse -- changes in prevalence of key content words related to AI and LLMs, but not necessarily the general themes or topics discussed in the student essays as identified through (dynamic) topic modeling.

John Edwards, Arto Hellas, Juho Leinonen

Computing educators and researchers have used programming process data to understand how programs are constructed and what sorts of problems students struggle with. Although such data shows promise for using it for feedback, fully automated programming process feedback systems have still been an under-explored area. The recent emergence of large language models (LLMs) have yielded additional opportunities for researchers in a wide variety of fields. LLMs are efficient at transforming content from one format to another, leveraging the body of knowledge they have been trained with in the process. In this article, we discuss opportunities of using LLMs for analyzing programming process data. To complement our discussion, we outline a case study where we have leveraged LLMs for automatically summarizing the programming process and for creating formative feedback on the programming process. Overall, our discussion and findings highlight that the computing education research and practice community is again one step closer to automating formative programming process-focused feedback.

Evanfiya Logacheva, Arto Hellas, James Prather et al.

Programming skills are typically developed through completing various hands-on exercises. Such programming problems can be contextualized to students' interests and cultural backgrounds. Prior research in educational psychology has demonstrated that context personalization of exercises stimulates learners' situational interests and positively affects their engagement. However, creating a varied and comprehensive set of programming exercises for students to practice on is a time-consuming and laborious task for computer science educators. Previous studies have shown that large language models can generate conceptually and contextually relevant programming exercises. Thus, they offer a possibility to automatically produce personalized programming problems to fit students' interests and needs. This article reports on a user study conducted in an elective introductory programming course that included contextually personalized programming exercises created with GPT-4. The quality of the exercises was evaluated by both the students and the authors. Additionally, this work investigated student attitudes towards the created exercises and their engagement with the system. The results demonstrate that the quality of exercises generated with GPT-4 was generally high. What is more, the course participants found them engaging and useful. This suggests that AI-generated programming problems can be a worthwhile addition to introductory programming courses, as they provide students with a practically unlimited pool of practice material tailored to their personal interests and educational needs.

Seth Bernstein, Paul Denny, Juho Leinonen et al.

Grasping complex computing concepts often poses a challenge for students who struggle to anchor these new ideas to familiar experiences and understandings. To help with this, a good analogy can bridge the gap between unfamiliar concepts and familiar ones, providing an engaging way to aid understanding. However, creating effective educational analogies is difficult even for experienced instructors. We investigate to what extent large language models (LLMs), specifically ChatGPT, can provide access to personally relevant analogies on demand. Focusing on recursion, a challenging threshold concept, we conducted an investigation analyzing the analogies generated by more than 350 first-year computing students. They were provided with a code snippet and tasked to generate their own recursion-based analogies using ChatGPT, optionally including personally relevant topics in their prompts. We observed a great deal of diversity in the analogies produced with student-prescribed topics, in contrast to the otherwise generic analogies, highlighting the value of student creativity when working with LLMs. Not only did students enjoy the activity and report an improved understanding of recursion, but they described more easily remembering analogies that were personally and culturally relevant.

Sami Sarsa, Juho Leinonen, Arto Hellas

We review and evaluate a body of deep learning knowledge tracing (DLKT) models with openly available and widely-used data sets, and with a novel data set of students learning to program. The evaluated knowledge tracing models include Vanilla-DKT, two Long Short-Term Memory Deep Knowledge Tracing (LSTM-DKT) variants, two Dynamic Key-Value Memory Network (DKVMN) variants, and Self-Attentive Knowledge Tracing (SAKT). As baselines, we evaluate simple non-learning models, logistic regression and Bayesian Knowledge Tracing (BKT). To evaluate how different aspects of DLKT models influence model performance, we test input and output layer variations found in the compared models that are independent of the main architectures. We study maximum attempt count options, including filtering out long attempt sequences, that have been implicitly and explicitly used in prior studies. We contrast the observed performance variations against variations from non-model properties such as randomness and hardware. Performance of models is assessed using multiple metrics, whereby we also contrast the impact of the choice of metric on model performance. The key contributions of this work are: Evidence that DLKT models generally outperform more traditional models, but not necessarily by much and not always; Evidence that even simple baselines with little to no predictive value may outperform DLKT models, especially in terms of accuracy -- highlighting importance of selecting proper baselines for comparison; Disambiguation of properties that affect performance in DLKT models including metric choice, input and output layer variations, common hyperparameters, random seeding and hardware; Discussion of issues in replicability when evaluating DLKT models, including discrepancies in prior reported results and methodology. Model implementations, evaluation code, and data are published as a part of this work.