Building computer science instruction for everyone requires understanding and supporting learner motivations beyond software development. Some learners value understanding the applications of code more than writing code from scratch (e.g., end-user programmers or conversational programmers). However, introductory programming instruction often emphasizes fundamental code-writing skills without clear applications. Our goal is to provide an overview of authentic computing applications to students from a variety of disciplines, including those who may not be interested in code-writing skills or may not have prior experience in programming. To this end, we designed and tested three interactive tutorials in the lab sessions of an introductory computing course for non-CS majors. Instead of teaching programming from scratch, these purpose-first tutorials highlight the purpose and basics of authentic programming libraries and tools, such as SQL queries, web-scraping libraries in Python, and HTTP requests. They are short (<1 hour), require minimal prior knowledge, and can be easily included in existing courses. The learning objectives of our activities are to help students understand why and when these tools are useful and to encourage them to explore these tools in problems that they might encounter in their disciplines. We observed that the activities were beneficial in improving confidence and providing a conceptual understanding of several tools. We present lessons learned to inform the design of similar activities for teaching computing as general knowledge to diverse audiences.

Pedagogical approaches focusing on stereotypical code solutions, known as programming plans, can increase problem-solving ability and motivate diverse learners. However, plan-focused pedagogies are rarely used beyond introductory programming. Our formative study (N=10 educators) showed that identifying plans is a tedious process. To advance plan-focused pedagogies in application-focused domains, we created an LLM-powered pipeline that automates the effortful parts of educators' plan identification process by providing use-case-driven program examples and candidate plans. In design workshops (N=7 educators), we identified design goals to maximize instructors' efficiency in plan identification by optimizing interaction with this LLM-generated content. Our resulting tool, PLAID, enables instructors to access a corpus of relevant programs to inspire plan identification, compare code snippets to assist plan refinement, and facilitates them in structuring code snippets into plans. We evaluated PLAID in a within-subjects user study (N=12 educators) and found that PLAID led to lower cognitive demand and increased productivity compared to the state-of-the-art. Educators found PLAID beneficial for generating instructional material. Thus, our findings suggest that human-in-the-loop approaches hold promise for supporting plan-focused pedagogies at scale.

Open-ended code-writing exercises are commonly used in large-scale introductory programming courses, as they can be autograded against test cases. However, code writing requires many skills at once, from planning out a solution to applying the intricacies of syntax. As autograding only evaluates code correctness, feedback addressing each of these skills separately cannot be provided. In this work, we explore methods to detect which high-level patterns (i.e. programming plans) have been used in a submission, so learners can receive feedback on planning skills even when their code is not completely correct. Our preliminary results show that LLMs with few-shot prompting can detect the use of programming plans in 95% of correct and 86% of partially correct submissions. Incorporating LLMs into grading of open-ended programming exercises can enable more fine-grained feedback to students, even in cases where their code does not compile due to other errors.

Knowledge tracing is a method to model students' knowledge and enable personalized education in many STEM disciplines such as mathematics and physics, but has so far still been a challenging task in computing disciplines.One key obstacle to successful knowledge tracing in computing education lies in the accurate extraction of knowledge components (KCs), since multiple intertwined KCs are practiced at the same time for programming problems.In this paper, we address the limitations of current methods and explore a hybrid approach for KC extraction, which combines automated code parsing with an expert-built ontology.We use an introductory (CS1) Java benchmark dataset to compare its KC extraction performance with the traditional extraction methods using a state-of-the-art evaluation approach based on learning curves.Our preliminary results show considerable improvement over traditional methods of student modeling.The results indicate the opportunity to improve automated KC extraction in CS education by incorporating expert knowledge into the process.

Background and Context: A major difference between expert and novice programmers is the ability to recognize and apply common and meaningful patterns in code. Previous works have attempted to identify these patterns as programming plans, such as counting or filtering the items of a collection. However, these efforts primarily relied on expert opinions and yielded many varied sets of plans. No methods have been applied to evaluate these various programming plans as far as their alignment with novices’ cognitive development.

Page-level analysis of documents has been a topic of interest in digitization efforts and multimodal approaches have been applied to both classification and page stream segmentation. In this work, we focus on capturing finer semantic relations between pages of a multi-page document. To this end, we formalize the task as semantic parsing of interpage relations and we propose an end-to-end approach for interpage dependency extraction, inspired by the dependency parsing literature. We further design a multi-task training approach to jointly optimize for page embeddings to be used in segmentation, classification, and parsing of the page dependencies using textual and visual features extracted from the pages. Moreover, we also combine the features from two modalities to obtain multimodal page embeddings. To the best of our knowledge, this is the first study to extract rich semantic interpage relations from multi-page documents. Our experimental results show that the proposed method increased LAS by 41 percentage points for semantic parsing, increased accuracy by 33 percentage points for page stream segmentation, and 45 percentage points for page classification over a naive baseline.