MilaNLP at MultiPRIDE: Evaluating Lexical, Transformer, and Retrieval-Augmented Models for Reclaimed Language
Feb 26, 2026·
,,·
0 min read
Ivana Crescenzi
Arianna Muti
Debora Nozza
Abstract
This paper describes our participation in the MultiPRIDE shared task at EVALITA 2026,
which focuses on the detection of reclaimed language within the LGBTQ+ community.
The system combines TF-IDF features with logistic regression and also evaluates
transformer-based models, instruction-tuned language models, and retrieval-augmented
approaches for reclaimed language detection.
Type
Publication
Proceedings of the Ninth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian. Final Workshop (EVALITA 2026)
Status
Peer-reviewed
Open access
Natural Language Processing
Reclaimed Language
Multilingual NLP
Large Language Models
Retrieval-Augmented Generation
Authors
PhD Student in Statistics and Computer Science
I am a PhD student in Statistics and Computer Science at Bocconi University.
My research interests include Natural Language Processing, Large Language Models,
trustworthy and personalized NLP, machine learning theory, algorithms and
optimization, with a particular interest in the theoretical analysis of
user-conditioned objective functions and the geometry and topology of
loss landscapes.