A summary of my professional activities

Sep 1, 2021: Starting a PhD in interpretable neural machine translation at the worderful GroNLP group under the supervision of Arianna Bisazza and Malvina Nissim, as a part of the NWO-funded consortium InDeep: Interpreting Deep Learning Models for Text and Sound.

Jul 31, 2021: Last day of work at Aindo. In the last months I worked on structured prediction from clinical reports, few-shot cross-lingual transfer for neural QA models, GNN for molecular properties prediction, neural recommender systems and sketch-to-image generation with GANs.

Jul 23, 2021: Participated in the HuggingFace JAX/Flax Community Week building the first CLIP image-text model for Italian (demo here) and the first T5 seq2seq model pre-trained on Italian. I also cleaned the largest available Italian corpus to date, the Italian split of mc4 (full size after cleaning ~103M docs, 41B words, 215Gb), making it available on the HuggingFace Hub.

June 6-11, 2021: Attended NAACL 2021, presenting That Looks Hard: Characterizing Linguistic Complexity in Humans and Language Models at CMCL 2021 - among best student papers! We also have our work Teaching NLP with Bracelets and Restaurant Menus: An Interactive Workshop for Italian Students out at TeachingNLP 2021.

Mar 1-3, 2021: Attended CLiC-it 2020 online. Alessio presented our work Italian Transformers Under the Linguistic Lens, that received a special mention for best student paper.

Jan 1, 2021: Started working full-time as ML research scientist for Aindo.

Dec 18, 2020: Attended EVALITA 2020 online, participating in the DANKMEMES and AcCompl-it tasks. My solo work UmBERTo-MTSA won a special mention among student papers!

Dec 11, 2020: Graduated cum laude from the Data Science MSc at the University of Trieste and SISSA with my thesis Interpreting Neural Langauge Models for Linguistic Complexity Assessment, under the supervision of Felice dell’Orletta and Davide Crepaldi.

Dec 1-3, 2020: Attended the 5th Workshop on Natural Language for AI (NL4AI 2020), co-located online at the AIxIA 2020 conference, presenting ETC-NLG: End-to-end Topic-Conditioned Natural Language Generation.

Nov 16-20, 2020: Attended EMNLP 2020 and CoNLL 2020.

Oct 19, 2020: Developed and open-sourcedLambdaBert, a BERT implementation based on lambda layers instead of self-attention. Viral on Made with ML.

Sep 28, 2020: Attended the “Deep Learning Theories in Neuroscience” session at the Oxford Virtual Autumn School in Neuroscience.

Jul 5-10, 2020: Attended ACL 2020 online.

May 8, 2020: Organizing the AI, Stats & COVID-19 event with AI2S, with more than 100 participants from around the world.

Apr 26-30, 2020: Attending the Eigth International Conference on Learning Representations (ICLR 2020) online.

Mar to Jun, 2020: Main developer for the Covid-19 Semantic Browser, a joint project by AILC and Area Science Park, covered in the news by many sources, including the NLP Newsletter by Sebastian Ruder and the Anthony Goldbloom talk at the Stanford HAI conference on COVID-19 and AI

Mar 22, 2020: Officially founded AI2S, the first AI student society in Friuli-Venezia Giulia.

Feb 13, 2020: Presented the “Hey Siri, what’s computational linguistics” interactive talk at SISSA Student Day 2020.

Feb 10-14, 2020: Attended the course “Eye-tracking Methods for Cognitive Science” by Elizabeth Schotter at SISSA.

Nov 12-15, 2019: Attending the Sixth Italian Conference on Computational Linguistics (CLiC-it 2019), in Bari, Italy.

Sep to Dec, 2019: Internship at the ItaliaNLP Lab of the Institute for Computational Linguistics ILC-CNR in Pisa. Worked under the supervision of Felice Dell’Orletta on linguistic complexity assessment with neural language models.

Sept 27-29, 2019: Presenting our interactive workshop “AItalo Svevo: Letters from an Artificial Intelligence” at Trieste Next 2019 in Trieste, Italy.

Jul 28 - Aug 2, 2019: Attending ACL 2019 in Florence, Italy.

May 8-10, 2019: Attending AILC’s Lectures on Computational Linguistics in Pavia, Italy, to present my work on topic modeling and sentiment analysis of the Svevo epistolary corpus.