Educational Natural Language Processing in Research and Practice

Educational Natural Language Processing in Research and Practice

📅 Thursday 21/05 16:00-19:00h

📍 Workshop Space B

🔎 Needs Analysis

Educational research and natural language processing (NLP) have long been mutually reinforcing. Recent advances in large language models have expanded educational and scientific applications. Language technologies are increasingly used for educational assessment (e.g., analyzing student responses and generating feedback), for studying classroom discourse and qualitative data, and for synthesizing scientific literature. At the same time, the growing volume of textual data makes manual analysis increasingly infeasible. Approaches such as scientific text mining, and retrieval-augmented generation provide important methodological building blocks for these tasks. This workshop equips participants with the foundations to apply language technologies critically and effectively in educational and research contexts, connecting core NLP methods with concrete uses cases from the educational domain.

📒 Session Description

The session will start with an introductory, interactive presentation, where we introduce the background and main methods (50min, shorter if the whole workshop is only accepted for 90min). Following this, we want the participants to work in groups of 3 or 4. The goal here is to conceptualize systems based on natural language processing techniques for their particular use case. We will propose example use cases such as feedback generation, literature synthesis, or educational recommender systems, but participants are free to also bring in their own ideas and use cases. The participants may use pen and paper to illustrate their ideas, and/or can work in Hyperchalk (https://hyperchalk.github.io/) boards.

💡 Learning Objectives

By the end of the workshop, participants will be able to:

1. Understand core NLP techniques for analyzing and generating educational and scientific texts.
2. Learn how language technologies can be applied to educational data such as student answers, classroom recordings, interviews, and forum discussions.
3. Learn how educational assessments can be scored using natural language processing techniques, and how we can generate according feedback.
4. Learn how scientific text mining and retrieval-augmented generation can support analytical, interpretive, and generative workflows.
5. Understand how extracted information can be structured and visualized as knowledge bases or knowledge graphs.