The VIABLE Lab joined the Festival of Learning 2026 in Seoul, Republic of Korea, with workshops and tutorials on June 27-28 and the main conference program from June 29-July 3, 2026. Four lab members - Dr. Anthony F. Botelho, Shan Zhang, Natalia S. Martin, and Hongming (Chip) Li - represented the lab on site. Seiyon M. Lee was unfortunately unable to join us in Seoul, and we missed having her there as part of the team.
The work presented extended well beyond the people who were able to travel. Zhongtian Huang and Dr. Huan (Hailey) Kuang made integral research contributions as co-authors of the AIED full papers, and each project reflected substantive partnerships with researchers across institutions. The conference week represented a larger collective effort than the group visible in any single photo.
The Festival brought the 27th International Conference on Artificial Intelligence in Education (AIED 2026), the 19th International Conference on Educational Data Mining (EDM 2026), and ACM Learning at Scale (L@S 2026) into one shared week at COEX. VIABLE Lab members contributed to five AIED main-track papers - four full papers and one short paper - plus one EDM poster, spanning technical, human, and methodological questions in AI and learning.
From Models to People at AIED 2026
Three full papers in the Technical Aspects of AIED track examined what it takes to make educational AI useful under real constraints. One study tested whether LLM self-reported confidence can support accountable dialogue coding and found that confidence must be calibrated and validated rather than accepted at face value. A second modeled variation in completion time across mathematics formative-assessment content, treating time not as background noise but as information about the relationship between learners and tasks. A third paper, from a separate external collaboration in which Shan Zhang participated, explored small, private language models as teammates in assessment design.
The Human Aspects of AIED contributions moved from model behavior to learner experience. Shan Zhang and collaborators used Ordered Network Analysis to study how middle school students' dialogue and behavior unfolded as they collaboratively designed AI chatbots. A related short paper examined an "attitude paradox": the relationship between students' ability beliefs and their intentions to persist in a conversational AI learning experience. Together, these studies ask not only whether an AI-supported activity works, but how learners interpret themselves, their peers, and the technology while participating in it.
EDM 2026: Help-Seeking Is Not One Behavior
At EDM, Shan Zhang and collaborators presented "Would a Short Explanation Immediately Help? Examining Causal Effects of Students' Help-Seeking in Math." The study combined Deep Knowledge Tracing with propensity-score weighting to estimate the effects of requesting on-demand explanations in ASSISTments. Average effects were not uniformly positive; the more informative result was heterogeneity. Students who spent more effort engaging with an explanation showed modest benefits relative to answer-only help, reinforcing that a support's value depends on how it is used, not simply whether it is available.
The poster session made that methodological point tangible. Conversations with WPI collaborators Eamon Worden and Neil Heffernan, and a detailed exchange with Professor Ken Koedinger, pushed the discussion beyond a single estimate toward better ways to distinguish sense-making from rapid answer-seeking in platform logs.
Leadership, Firsts, and Long-Running Partnerships
Research presentations were only one part of the lab's participation. Dr. Botelho served as an EDM 2026 Program Chair, helping shape a program centered on connections across disciplines, contexts, and research communities. Chip Li served as an EDM 2026 Web Chair, supporting the conference website and its online information throughout the event.
Shan Zhang also co-organized two full-day Festival workshops on June 28. The 10th Educational Data Mining in Computer Science Education (CSEDM) Workshop connected EDM and computing education researchers around AI- and data-driven approaches to understanding how students learn computer science. The AI Literacy For All: 2nd International Workshop on AI Literacy Education For All brought together research and practice across K-12, higher education, educators, and workforce learning. These organizing roles extended the lab's participation from presenting research to helping create spaces for broader community exchange.
The week also marked Natalia S. Martin's first international academic conference. Presenting an accepted workshop poster in Seoul was an important step from joining a research community to actively contributing to it. Across the lab, the week reflected several kinds of growth at once: publishing new work, taking on service and leadership, and learning how to explain ideas to audiences with very different methodological backgrounds.
Presented Work
AIED 2026 - Technical Aspects of AIED, Full Papers
[1] Can We Trust AI's Self-Assessment? Evaluating and Improving LLM Confidence Calibration in Educational Dialogue Coding
Li, H., Kuang, H., & Botelho, A. F. (2026). DOI: 10.1007/978-3-032-29755-6_18 | Detailed project page and talk materials
[2] Small, Private Language Models as Teammates for Educational Assessment Design
Jaldi, C. D., Saini, A., Zhang, S., Schroeder, N., Shimizu, C., & Ilkou, E. (2026). DOI: 10.1007/978-3-032-29755-6_15
External collaboration involving Shan Zhang.
[3] Modeling Completion Time in Mathematics Formative Assessments: Content-Based Prediction of Time Variation
Botelho, A. F., Huang, Z., Martin, N. S., & Shin, J. (2026). DOI: 10.1007/978-3-032-29755-6_21
AIED 2026 - Human Aspects of AIED
[4] Full Paper: Analyzing Middle School Students' Dialogue and Behaviors during Collaborative AI Chatbot Development Using Ordered Network Analysis
Zhang, S., Zambrano, A. F., Tian, X., Song, Y., Botelho, A. F., Boyer, K. E., Israel, M., & Jiang, S. (2026). DOI: 10.1007/978-3-032-29763-1_23
[5] Short Paper: An Attitude Paradox? Examining Ability Beliefs and Persistence Intentions in a Middle School Conversational AI Learning Experience
Tian, X., Zhang, S., Song, Y., McKlin, T., Boyer, K. E., & Israel, M. (2026). DOI: 10.1007/978-3-032-29770-9_57
EDM 2026 - Poster and Demo Track
[6] Would a Short Explanation Immediately Help? Examining Causal Effects of Students' Help-Seeking in Math
Zhang, S., Worden, E., Leite, W., Heffernan, N. T., & Botelho, A. F. (2026). DOI: 10.5281/zenodo.21039683 | EDM proceedings
Seoul Beyond COEX
Seoul gave the week a different rhythm from the conference room. COEX placed the Festival in the middle of Gangnam's constant motion, while evenings opened into Myeongdong food stalls, neon-lit streets, and conversations that were easier to continue without a slide timer running.
Some of the best moments were deliberately simple: a Han River picnic with friends and longtime collaborators from WPI, followed on another evening by a karaoke outing that traded research questions for a shared song queue. The week strengthened collaborations because people had time to be people together - a fitting ending for a Festival built around learning across communities.