PhD Student at the University of Milano-Bicocca.
Interested in Artificial Intelligence, Natural Language Processing, Machine Learning and Hyperbolic Geometry
@xanamini@grok which are other tech accounts that posted april fools like this and became popular yesterday? link me those, or if you can't write the account names
I was finally able to sit down to write a blog post about our experience in training CLIP-Italian during the wonderful @huggingface community week :)
Link: https://t.co/jcnAGe1REX
🎉 #EMNLP2021 new paper! "SWEAT: Scoring Polarization of Topics across Different Corpora" with @MatteoPalmonari @MarelliMar and P. Nicoli.
SWEAT shows if two corpora hold polarized views about a topic (e.g. politics, parenting), reporting effect size and significance.🧵
😻🚀💥New paper with @dirk_hovy accepted to the Findings of ACL2021! #ACL2021NLP#NLProc
"On The Gap between Understanding and Adoption in NLP": we describe several issues that can hamper scientific development in our community.
pre-print out soon!
Thrilled to share that our paper on topic modeling has been accepted to #ACL2021NLP! w/ @TerragniSilvia and @dirk_hovy.
Adding contextualized representations improves the coherence of neural topic models without affecting the diversity!
Python package: https://t.co/f8ej6LMfJn
"Query2Prod2Vec: Grounded Word Embeddings for eCommerce" accepted at #NAACL2021 Industry Track! with @jacopotagliabue and @christineyyuu (@coveo).
Query2Prod2Vec grounds lexical representations for product search in product embeddings and achieves SOTA performance!
#NLProc
Vuoi applicare algoritmi di #Sentiment e #Emotion analysis in Italiano?
Abbiamo appena rilasciato una nuova libreria Python https://t.co/pixifUjohZ
Oltre a due language models su @huggingface
👉emotion: https://t.co/J6HplnI1oP
👉sentiment: https://t.co/0zYSgy8feI
#NLProc
Paper accepted at #NAACL2021: "Language in a (Search) Box: Grounding Language Learning in Real-World Human-Machine Interaction” with @jacopotagliabue and @GreCo_CiRo (@coveo).
We show a way of learning meaning from scratch using search interactions! #NLProc
Paper summary: 1/6
🎉🎉Our paper "Cross-lingual Contextualized Topic Models with Zero-shot Learning" has been accepted at #EACL2021.
Paper: https://t.co/hcZbpgHs0a (preprint)
Python package: https://t.co/f8ej6LMfJn
w/ @TerragniSilvia @dirk_hovy @debora_nozza@FersiniE#NLProc#ai 1/6
Happy to share our new pre-print: "BERT Goes Shopping: Comparing Distributional Models for Product Representations" with @christineyyuu and @jacopotagliabue
We introduce ProdBERT for eCommerce product representations.
pre-print: https://t.co/B1tCURnJqa
#NLProc#ecommerce#ai
Interested in comparing word meaning differences between the #Pokémon world and science? We just released CADE a framework to align and compare word embeddings across corpora.
Paper: https://t.co/hL66qkI5ro (generalization of our #AAAI2019 paper).
Code: https://t.co/c3kyxRNKKr
Do you fear you have lost a BERT-based joke? we got you covered! we have gathered them in one website!
Next time someone says CamemBERT at a party you can laugh (this time for real!). See our table, with models and tasks at https://t.co/moJGzndFuR