FFT-UniBa at Cruciverb-IT: Special Length Tokens and CSP for Italian Crossword Solving
An Italian crossword-solving system combining an Italian T5 model with special answer-length tokens and a constraint satisfaction solver to fill complete grids.
A little introduction
Welcome to my corner of the web.
I am a PhD student at CILab, in the Department of Computer Science at the University of Bari Aldo Moro. My research interests focus on large language models, mechanistic interpretability, multimodal AI and trustworthy AI.
More about meSelected publications
All publicationsTraining Green and Sustainable Recommendation Models: Introducing Carbon Footprint Data into Early Stopping Criteria
A carbon-aware early stopping strategy for recommender systems that weighs improvements in predictive accuracy against training emissions, reducing the environmental footprint with limited changes in accuracy and fairness.