Background: Context personalization, the practice of adapting learning materials to students’ personal interests, has been shown to increase student learning and engagement. Within computer science education, research has found that LLMs can generate high-quality contextualized introductory programming exercises. Objective: In this paper, we evaluate the capability of LLMs to generate technically correct and thematically integrated contextualized algorithm design problems. Methods: In a series of three iterative studies, we use LLMs to generate contextualized algorithm design problems from a given base problem and theme, evaluating over 500 generated problems for technical and thematic alignment. Results: We find that LLM-generated algorithm design problems exhibit significantly more issues than prior work has found for introductory programming problems. We identify issues specific to the algorithm design context and then mitigate these issues with prompt engineering techniques and model choice. With these adjustments, we produce LLM-generated contextualized algorithm design problems that are technically strong, deeply themed, and largely realistic, though realism drops with more culturally and locally specific themes. Implications: We demonstrate a viable workflow for generating contextualized algorithm design problems using LLMs, including prompt design, model selection, and identification of specific issues to review for.

The ability to identify the important takeaways from a previously-seen solution and apply them in different contexts is an important problem-solving skill. However, this skill, known as analogical reasoning, is traditionally left implicit in algorithms courses. Students are expected to develop the skill naturally as they progress through the course. In this study, we aim to conduct a more thorough investigation of analogical reasoning in algorithms. We integrate explicit metacognitive scaffolds for reflection and schema development into an undergraduate algorithms course. Then, on course exams, we insert an additional task alongside select algorithm design questions, in which students are asked to describe how a previously-seen problem influenced their design. We analyzed both the previously-seen problem selected by the student and the stated similarity. Within the 142 comparisons analyzed, we find that 37% provide insight about the underlying solution structure, and these comparisons were significantly associated with higher scores on the problem. Furthermore, about one-third of the comparisons were with a problem that course staff also selected, and these comparisons were not only much more likely to be structural but were also correlated with higher performance on the question. Our results indicate that the analogical reasoning skills are closely tied to success in the algorithms course, and encourage instructors to integrate explicit demonstrations into their curriculum.

Dynamic Programming (DP) is commonly regarded as one of the most difficult topics in the upper-level algorithms curriculum. The teaching of metacognitive strategies may prove effective in helping students learn to design DP algorithms. To explore both whether students learn and use these strategies on their own and the effect of guidance about using these strategies, we conducted think-aloud interviews with structured guidance at two points in a college algorithms course: once immediately after students learned the concept and once at the end of the course. We explore 1) what metacognitive strategies are commonly employed by students, 2) how effectively they help students solve problems, and 3) to what extent structured guidance about using metacognitive strategies is effective. We find that these strategies generally help students make progress in solving DP problems, but that they can mislead students as well. We also find that the adoption of these strategies is an individualized process and that structured strategy guidance is often insufficient in allowing students to solve individual DP problems, indicating the need for more extensive strategy instruction.

According to an ecological affordances perspective, any static curriculum has a set of affordances, and differences in teachers, students, and the teaching environment change how those affordances are viewed and used. Therefore, teaching is a relationship between the curriculum, the teacher, and the students. As such, it is not only possible but expected that a teacher will diverge from the details of a lesson plan to better accommodate the needs of themselves as a teacher and their students as learners.

Dynamic programming (DP) is one of the most challenging topics in algorithms courses. Although there exist animation tools that assist with the understanding of DP algorithms, few existing tools are aimed at scaffolding the process of solving DP algorithm design problems. To help create a learning tool able to provide the affordances learners need when attempting DP problems, we analyzed learners' scratch paper to understand how learners approach DP problems. Based on scratch paper from 18 learners solving DP problems during a think-aloud study, we created a codebook that characterized different elements and methods used by the learners on their scratch paper. We found that learners had distinct preferences when attempting DP problems. Some learners preferred using example input with specific values to simulate ideal program executions, while some used math representations of example inputs to help derive formulas. Learners interacted with their example input in multiple ways, including filling in hand-drawn tables to organize the calculation process and dynamically interacting with the inputs by crossing, circling, or using arrows to visualize the relationships between inputs. These findings suggest potential interactions that need to be taken into consideration when designing tools to support learners in solving DP problems.

Algorithm design is a vital skill developed in most undergraduate Computer Science (CS) programs, but few research studies focus on pedagogy related to Algorithms coursework. To understand the work that has been done in the area, we present a systematic survey and characterization of existing studies in the CS Education literature related to the teaching of algorithm design at the undergraduate level. Across all papers in the ACM Digital Library, we only find 97 applicable papers. We classify these papers by topic, evaluation metric, evaluation methods, and intervention target. We present the results of these classifications alongside insights about existing knowledge, rigor, and contribution rates. We hope that this work not only provides a detailed representation of the current corpus of CS Education work related to algorithm design but also demonstrates that the body of knowledge is sparse and supports further research in the area. For future work, we intend to investigate and synthesize the conclusions reached by these papers.

Computer science learning in primary school classrooms has expanded, necessitating effective instructional strategies for this age group. Gallery Walks are a common activity to allow peers to share their work and give feedback on peers’ work. Like other skills, providing effective feedback may require scaffolding and/or instruction for some students.

Quantum computing presents a paradigmatic shift in the field of computation, in which unintuitive properties of quantum mechanics can be harnessed to change the way we approach a wide range of problems. However, due to the mathematics and physics perspective through which quantum computing is traditionally presented, most resources are inaccessible to many undergraduate students, let alone the general public. It is thus imperative to develop resources and best-practices for quantum computing instruction accessible to students at all levels. In this paper, we describe the development and results of our Massive Open Online Course (MOOC) "Introduction to Quantum Computing for Everyone." This course presents an introduction to quantum computing with few technical prerequisites. In the first half of the course, quantum computing concepts are introduced with a unique, purely visual representation, allowing students to develop conceptual understanding without the burden of learning new mathematical notation. In the second half, students are taught the formal notation for concepts and objects already introduced, reinforcing student understanding of these concepts and providing an applicable context for the technical material. Most notably, we find that introducing the math content in the curriculum's second stage led to no drops in engagement or student performance, suggesting that our curriculum's spiral structure eased the technical burden.

We present the CS1 Reviewer App - an online tool for an introductory Python course that allows students to solve customized problem sets on many concepts in the course. Currently, the app's questions focus on code tracing by presenting a block of Python code and asking students to predict the output of the code. The tool tracks a student's response history to maintain a "mastery level" that represents a student's knowledge of a concept. We also provide an option of answering auto-generated quizzes based on the student's mastery across concepts. As a result, the tool provides students a choice between creating their own learning experience or leveraging our question selection algorithm. The app is supported on traditional webpages and mobile devices, providing a convenient way for students to study a variety of concepts. Students in the CS1 course at Duke University used this tool during the Spring and Fall 2020 semesters. In this paper, we explore trends in usage, feedback and suggestions from students, and avenues of future work based on student experiences.

The eukaryotic transcription cycle consists of three main steps: initiation, elongation, and cleavage of the nascent RNA transcript. Although each of these steps can be regulated as well as coupled with each other, their in vivo dissection has remained challenging because available experimental readouts lack sufficient spatiotemporal resolution to separate the contributions from each of these steps. Here, we describe a novel application of Bayesian inference techniques to simultaneously infer the effective parameters of the transcription cycle in real time and at the single-cell level using a two-color MS2/PP7 reporter gene and the developing fruit fly embryo as a case study. Our method enables detailed investigations into cell-to-cell variability in transcription-cycle parameters as well as single-cell correlations between these parameters. These measurements, combined with theoretical modeling, suggest a substantial variability in the elongation rate of individual RNA polymerase molecules. We further illustrate the power of this technique by uncovering a novel mechanistic connection between RNA polymerase density and nascent RNA cleavage efficiency. Thus, our approach makes it possible to shed light on the regulatory mechanisms in play during each step of the transcription cycle in individual, living cells at high spatiotemporal resolution.

Fusing of multi-modal data involves automatically estimating the coordinate transformation required to align the multi-modal image data sets. Most existing methods in literature are not fast enough (take hours for estimating nonrigid deformations) for practical use. We propose a very fast algorithm, based on matching local-frequency image representations, which naturally allows for processing the data at different scales/resolutions, a very desirable property from a computational efficiency view point. This algorithm involves minimizing-over all affine transformations-the expectation of the squared difference between the local-frequency representations of the source and target images. In cases where fusing the multi-modal data requires estimating the non-rigid deformations, we propose a novel and fast PDE-based morphing technique that will estimate this non-rigid alignment. We present implementation results for synthesized and real misalignments between CT and MR brain scans. In both the cases, we validate our results against ground truth registrations which for the former case are known and for the latter are obtained from manual registration performed by an expert.