COST Action Training School in Computational Opinion Analysis – COpA 3rd Ed.

August 26-27, 2026

Wednesday:

08:30 – 09:00: Registration Participant check-in for the event.
09:00 – 09:45: Presentation of the Training School & Conceptual Introduction General introduction to the purpose of the event and its conceptual framework. Ethical framework. Carlos Arcila Calderón & David Blanco
09:45 – 10:00: Break.
10:00 – 12:15: LLMs for Opinion Analysis. Using LLMs as annotators or classifiers. Ana Jovancevic
12:15 – 13:30: Lunch
13:30 – 14:30: LLMs for Opinion Analysis. A focus on biases, hallucinations, reproducibility. Aleksandra Urman
14:30 - 14:45: Break.
14:45 – 16:30: Validation and interpretability. Related to validation, biases, explainability. Aleksandra Urman
17:00: Social event

Thursday:

10:00 – 11:00: Comparative and cross-context analysis. Inspiring application cases of multi-lingual and multi-modal analysis. Christian Baden
11:00 – 11:15: Break.
11:15 – 13:15: Annotation, gold standards, Disagreement. Types of annotation (crowd vs experts), dealing with ambiguity, intercoder reliability. Mariken van der Velden
13:15 – 14:15: Lunch.
14:15 – 15:45: Fine-tuning. Fine-tuning LLMs. Ljubisa Bojic
15:45 – 16:00: Break
16:00 – 17:30: Model evaluation. Strategies to evaluate traditional and LLM models, generalization of findings. Ljubisa Bojic
17:30 – 18:00: Closing Final summary, reflections on what was learned, and formal closing of the event. How to teach and transfer this knowledge. Carlos Arcila Calderón & David Blanco