Lujie Chen

Md Biplob Hosen, Md Alomgeer Hussein, Md Akmol Masud et al.

Patient portals now give individuals direct access to their electronic health records (EHRs), yet access alone does not ensure patients understand or act on the complex clinical information contained in these records. The ArchEHR-QA 2026 shared task addresses this challenge by focusing on grounded question answering over EHRs, and this paper presents the system developed by the HealthNLP_Retrievers team for this task. The proposed approach uses a multi-stage cascaded pipeline powered by the Gemini 2.5 Pro large language model to interpret patient-authored questions and retrieve relevant evidence from lengthy clinical notes. Our architecture comprises four integrated modules: (1) a few-shot query reformulation unit which summarizes verbose patient queries; (2) a heuristic-based evidence scorer which ranks clinical sentences to prioritize recall; (3) a grounded response generator which synthesizes professional-caliber answers restricted strictly to identified evidence; and (4) a high-precision many-to-many alignment framework which links generated answers to supporting clinical sentences. This cascaded approach achieved competitive results. Across the individual tracks, the system ranked 1st in question interpretation, 5th in answer generation, 7th in evidence identification, and 9th in answer-evidence alignment. These results show that integrating large language models within a structured multi-stage pipeline improves grounding, precision, and the professional quality of patient-oriented health communication. To support reproducibility, our source code is publicly available in our GitHub repository

Varun Mandalapu, Lujie Karen Chen, Sushruta Shetty et al.

In recent years, there is a lot of interest in modeling students' digital traces in Learning Management System (LMS) to understand students' learning behavior patterns including aspects of meta-cognition and self-regulation, with the ultimate goal to turn those insights into actionable information to support students to improve their learning outcomes. In achieving this goal, however, there are two main issues that need to be addressed given the existing literature. Firstly, most of the current work is course-centered (i.e. models are built from data for a specific course) rather than student-centered; secondly, a vast majority of the models are correlational rather than causal. Those issues make it challenging to identify the most promising actionable factors for intervention at the student level where most of the campus-wide academic support is designed for. In this paper, we explored a student-centric analytical framework for LMS activity data that can provide not only correlational but causal insights mined from observational data. We demonstrated this approach using a dataset of 1651 computing major students at a public university in the US during one semester in the Fall of 2019. This dataset includes students' fine-grained LMS interaction logs and administrative data, e.g. demographics and academic performance. In addition, we expand the repository of LMS behavior indicators to include those that can characterize the time-of-the-day of login (e.g. chronotype). Our analysis showed that student login volume, compared with other login behavior indicators, is both strongly correlated and causally linked to student academic performance, especially among students with low academic performance. We envision that those insights will provide convincing evidence for college student support groups to launch student-centered and targeted interventions that are effective and scalable.

Varun Mandalapu, Jiaqi Gong, Lujie Chen

Interactive Educational Systems (IES) enabled researchers to trace student knowledge in different skills and provide recommendations for a better learning path. To estimate the student knowledge and further predict their future performance, the interest in utilizing the student interaction data captured by IES to develop learner performance models is increasing rapidly. Moreover, with the advances in computing systems, the amount of data captured by these IES systems is also increasing that enables deep learning models to compete with traditional logistic models and Markov processes. However, it is still not empirically evident if these deep models outperform traditional models on the current scale of datasets with millions of student interactions. In this work, we adopt EdNet, the largest student interaction dataset publicly available in the education domain, to understand how accurately both deep and traditional models predict future student performances. Our work observes that logistic regression models with carefully engineered features outperformed deep models from extensive experimentation. We follow this analysis with interpretation studies based on Locally Interpretable Model-agnostic Explanation (LIME) to understand the impact of various features on best performing model pre-dictions.