Students AY 2024-2025

Final session. We discussed IAA findings, introduced the idea of building a custom GPT, and reviewed the final project and exam format.

Presentation of updated results and materials. A revised version of the tagset and manual was introduced to address inconsistencies. We also discussed implications of GPT-assisted annotation.

Preliminary IAA results were shared. We introduced a small experiment with GPT-assisted annotation: students reviewed and corrected ChatGPT’s output. Simple error category analyses using R were also introduced.

Focus on IAA: students annotated three interactions independently. The aim was to assess consistency across annotators and reflect on differences in tagging.

Continued work on shared annotation tasks. Students applied the Louvain tagset to new texts in preparation for the inter-annotator agreement (IAA) phase.

First feedback on metadata annotation. Introduction to error annotation using the Louvain taxonomy and the UCLEE software. Students annotated a sample text and used shared documents to resolve tagging issues collaboratively.

Discussion of chatbots in language learning and the role of learner corpora. Students selected texts from the UNITE project and began annotating them with learner- and task-related metadata.

Introductory session outlining the course aims and structure, with a brief overview of learner corpus research and its relevance to second language acquisition.

• Cecilia Brighittini
• Jennifer Di Bonaventura
• Angela Forzatti
• Joanna Giacobbe
• Anastasiya Olegovna Khan
• Marta Salvatore
• Davide Pio Santangelo
• Martina Simonotti
• Federica Spolverini
• Maura Volpe

• Adriano Ferraresi
• Giada Palmieri
• Daniele Polizzi