The growing adoption of Large Language Models (LLMs) is transforming learning in higher education. While they provide flexible support, concerns remain about over-reliance, uncritical acceptance of outputs, and excessive cognitive offloading. To address these challenges, we foreground self-regulated learning for LLMs (SRL-for-LLM), defined as learners' ability to plan, monitor, and reflect on AI interactions to support learning. Guided by SRL theory, we introduce ChatWise, a proof-of-concept browser extension that combines conversational scaffolding with learning analytics to classify student prompts by SRL strategies and deliver adaptive metacognitive feedback in real time. A two-phase study included the design of ChatWise and a within-subject field study with STEM undergraduates. Log analyses using mixed-effects modeling showed that access to ChatWise more than doubled the likelihood of producing high-quality prompts. Findings further indicate improved prompt refinement, greater strategic awareness, and more reflective engagement with AI-supported learning. These results highlight the potential of SRL-aligned, analytics-driven scaffolding to support more effective student–LLM interactions.

Jupyter is a web-based, interactive computing environment that supports many commonly-used programming languages. It has been widely adopted in the CS education community and is now rapidly expanding to other STEM disciplines due to the growing integration of programming in STEM education. However, unlike other educational platforms, there is currently no integrated way to capture, analyze, and visualize student interaction data in Jupyter notebooks. This means that teachers have limited to no visibility into student activity, preventing them from drawing insights from these data and providing timely interventions on the fly. In this paper, we present Jupyter Analytics, an end-to-end solution for teachers to collect, analyze, and visualize both synchronous and asynchronous learning activities in Jupyter. The Jupyter Analytics system consists of two JupyterLab extensions connected via a cloud-based backend. On the student side, we introduce the Jupyter Analytics Telemetry extension to anonymously capture students' interaction activity with more structure and higher granularity than log data. On the teacher side, we introduce the Jupyter Analytics Dashboard extension, which visualizes real-time student data directly in the notebook interface. The Jupyter Analytics system was developed through an iterative co-design process with university instructors and teaching assistants, and has been implemented and tested in several university STEM courses. We report two use cases where Jupyter Analytics impacted teaching and learning in the context of exercise sessions, and discuss the potential value of our tools for CS education.

This paper investigates the potential impact of deep generative models on the work of creative professionals. We argue that current generative modeling tools lack critical features that would make them useful creativity support tools, and introduce our own tool, generative.fashion1, which was designed with theoretical principles of design space exploration in mind. Through qualitative studies with fashion design apprentices, we demonstrate how generative.fashion supported both divergent and convergent thinking, and compare it with a state-of-the-art text-based interface using Stable Diffusion. In general, the apprentices preferred generative.fashion, citing the features explicitly designed to support ideation. In two follow-up studies, we provide quantitative results that support and expand on these insights. We conclude that text-only prompts in existing models restrict creative exploration, especially for novices. Our work demonstrates that interfaces which are theoretically aligned with principles of design space exploration are essential for unlocking the full creative potential of generative AI.

How to measure students' Computational Problem-Solving (CPS) competencies is an ongoing research topic. Prevalent approaches vary by measurement tools (e.g., interactive programming, multiple-choice tests, or programming-independent tests) and task types (e.g., debugging problems or Parson problems). However, few studies have examined the measurement tools of CPS competencies themselves: affordances and limitations of the measurement tools and how they compare, or whether different task types might elicit CPS competencies differently. Research needs to address these questions in order to better understand how to design robust, generalizable, and effective measurement tools for CPS competencies. This paper presents an exploratory study that contributes to this research direction. It is part of a larger international project to develop an open-access formative assessment platform for CPS, which includes a novel authoring tool for a wide range of task types for interactive block-based programming. We used the tool to create an interactive programming experience with multiple task types and gave it to more than 300 secondary school students from different countries. We also administered a validated multiple-choice measurement of Computational Thinking with block-based programs. We focused on task complexity as a characteristic of task type, using a classification scheme based on task design features. Comparing students' performances on tasks of different complexity and using two distinct measurement tools, we found that the multiple-choice measurement only partially predicts performance in the interactive programming task. Additionally, its predictive capacity varies significantly between task types of differing complexity.

Algorithms deployed in education can shape the learning experience and success of a student. It is therefore important to understand whether and how such algorithms might create inequalities or amplify existing biases. In this paper, we analyze the fairness of models which use behavioral data to identify at-risk students and suggest two novel pre-processing approaches for bias mitigation. Based on the concept of intersectionality, the first approach involves intelligent oversampling on combinations of demographic attributes. The second approach does not require any knowledge of demographic attributes and is based on the assumption that such attributes are a (noisy) proxy for student behavior. We hence propose to directly oversample different types of behaviors identified in a cluster analysis. We evaluate our approaches on data from (i) an open-ended learning environment and (ii) a flipped classroom course. Our results show that both approaches can mitigate model bias. Directly oversampling on behavior is a valuable alternative, when demographic metadata is not available. Source code and extended results are provided in https://github.com/epfl-ml4ed/behavioral-oversampling.

The ability to consider a wide range of solutions to a design problem is a crucial skill for designers, and is a major differentiator between experts and novices. One reason for this is that novices are unaware of the full extent of the design space in which solutions are situated. To support novice designers with design space exploration, we introduce Mixplorer, a system that allows designers to take an initial design and mix it with other designs. Mixplorer differs from existing tools by supporting the exploration of ill-defined design spaces through social design space exploration. To evaluate Mixplorer, we conducted (1) an interview study with design instructors who reported that Mixplorer would “help to open the minds” of novice designers and (2) a controlled experiment with novices, finding that the design-mixing functionality of Mixplorer provided significantly better support for creativity, and that participants who mixed designs produced more novel designs.

Haptic force feedback systems are unique in their ability to dynamically render physical representations. Although haptic devices have shown promise for supporting learning, prior work mainly describes results of haptic-supported learning without identifying underlying learning mechanisms. To this end, we designed a haptic-supported learning environment and analyzed four students who used it to make connections between two different mathematical representations of sine and cosine: the unit circle, and their graph on the Cartesian plane. We highlight moments where students made connections between the representations, and identify how the haptic feedback supported these moments of insight. We use this evidence in support of a proposed theoretical and design framework for educational haptics. This framework captures four types of haptic representations, and focuses on one -- the haptic bridge -- that effectively scaffolds sense-making with multiple representations.

This paper introduces a new environment for programming robots and physical computing devices---the Spatial Computing Platform (SCP)---and compares it to a text-based programming environment (the Cricket Logo). The SCP simplifies the process of constructing conditional statements that link the robot's inputs and outputs together. It does this by providing the user with a virtual canvas that they can draw rectangles on using the mouse. Each rectangle represents a range of sensor values, and specific outputs can be assigned to each rectangle. When the sensor values enter into the specified range, the outputs will turn on. We designed a study with 60 youth to compare this environment to Cricket Logo, a well-known variant of Logo designed to control robotic devices. We found that participants using the spatial computing platform were able to build programs of higher complexity and make more changes to their programs over the course of an hour-long workshop.

Connected Messages brings together traditions of engaging youth in designing interactive murals with themes relevant to their lives and new lowcost networking technologies of connecting local groups with global audiences. We describe the design of a community mural that functions like a public display, which can be remotely programmed through an online interface. The implementation with Maker Mentors in five Free Library branches with over 1,000 youth focused on different aspects of youth maker agency in accessing, participating, and expressing their ideas. In the discussion, we review key dimensions expanding youth and Maker Mentor participation in communityrelevant designs.

The recent development of low cost tangible construction kits has lowered the barriers to entry for amateur and youth hardware designers. In this paper, we discuss the outcomes of a pilot study in which middle school youths set up a game arcade by remixing Scratch games and crafting physical game controllers using the MaKey MaKey and Play-Doh. The analyses focused on the youths' designs and reflections on Scratch remixes, their game controllers, and the culminating arcade. Designing game controllers and setting up the arcade was a productive learning experience for the youths in which they were able to glean insights about how to improve their designs by watching younger students play their games. In the discussion we highlight the parts of the workshop that were most successful, address what we learned about the youths' experiences, develop suggestions for the design of future workshops, and discuss the significance of authentic audience designs.