Emergence of a high-dimensional abstraction phase in language transformers. Proceedings of ICLR 2025 (International Conference on Learning Representations), online at https://openreview.net/group?id=ICLR.cc/2025/Conference. M. Mahaut, F. Franzon, R. Dessi and M. …

Welcome to my homepage! I am an ICREA research professor, affiliated with the Department of Linguistics at Pompeu Fabra University, in Barcelona (here is my ICREA page). I study natural and artificial intelligence, with a focus on language-mediated intelligence. I am currently heading the ERC-funded ALiEN project on the emergence of communication in populations of deep networks. You can refer ...

Co-organizer and coordinator of the BA and MA majors in Philosophy and Informatics of the Philosophy program at the University of Trento, 2009-2013 Semantics area co-chair at EMNLP 2015 (Conference on Empirical Methods in Natural Language Processing), September 2015 External evaluator of Marco Turchi's tenure track at FBK (Trento), 2013-2015

Knorpora: the Knoppix Linux live CD remastered with tools and resources for corpus-based computational linguistics students; English token and document frequency lists from the LOB and Brown corpora; regexp_tokenizer: a simple tokenizer based on regular expressions. Back to …

This page contains information and resources pertaining to the Distributional Memory approach to corpus-based semantics proposed in the following article: M. Baroni and A. Lenci. Distributional Memory: A general framework for corpus-based semantics. …

M. Marelli, S. Menini, M. Baroni, L. Bentivogli, R. Bernardi and R. Zamparelli (2014). A SICK cure for the evaluation of compositional distributional semantic models.

Unnatural language processing Kevardec, Franzon and Baroni Findings of EMNLP 2023

not sophisticated enough to cope with the needs of linguists (Baroni and Bernardini 2007), and chances are slim that they will ever be, thus making the provision of very large and up-to-date …

The Representation of Prefixed Forms in the Italian Lexicon: Evidence from the * Distribution of Intervocalic [s] and [z] in Northern Italian Marco Baroni, UCLA

Compositional languages are “viral”! Are compositional languages always the easiest to spread? No! This will depend on the input (Lazaridouet al ICLR 2018), the task, and what comes easy …