Ciao a tutti!
Following the example of other language communities, this is the introduction thread for Italian NLP practitioners using HuggingFace libraries! To get started, please feel free to introduce yourself with any of the following:
- Your name, Github, Hugging Face, and/or Twitter handle
- Some projects you are working on or interested in starting
- Any potential directions you have in mind for the Hugging Face Italian community
- Anything else you’d like to share!
As for me, I am Gabriele (@gsarti on Github, gsarti_ on Twitter), I recently graduated from a Data Science MSc program at University of Trieste/SISSA and I’m currently working for an AI startup + looking for NLP PhD opportunities!
As for my recent projects using HF libraries, I recently used the
allenai/unifiedqa-t5-large model alongside the new prefix-constrained generation capabilities introduced by @nicola-decao for multiple choice QA!
Looking forward to e-meet you all!
good idea to open the thread, we need to strengthen the nlp community in Italy
I’m Enrico Bertino, mathematical engineer passionate about language (@EnricoBerto91 on twitter).
I co-founded Indigo.ai, a startup in Ai with a focus on R&D in the NLP field.
Speaking about HF, I use transformers in practically all the task I handle. In Indigo, we love “light” models and we pre-trained distilBERT in Italian (it was the only chance in my life to call a model like me ) and this year we aim to train some even more compressed models in order to allow the use of transformers in Italian also with low computational demand.
In parallel with BERT, we are also exploring the world of “next word predictor” language models, currently training a GPT-2 and having fun with various experiments with GPT-3! With this in mind, it would be interesting to organize a discussion on GPT-3 and its use in Italian, I think there are ideas from both a technical and social point of view!
In ogni caso sono super carico per fare una comunità HF italiana!!
Ciao Gabriele, bellissima idea!
I am Cristiano De Nobili. After a Ph.D. in Theoretical Physics @ SISSA (Trieste) and a short Postdoc in deep learning (computer vision), I have been into NLP for three years. I started working at the Bixby project (Samsung’s vocal assistant) and now I am a freelance research scientist mainly working on NLP/DL projects that have a positive social or environmental impact. Currently, together with Pi School and the EU commission, I am working to apply NLP to have a better understanding of industrial pollution in Europe. In addition, I am a deep learning lecturer at several public and private masters, in particular at the master in High-Performance Computing MHPC (SISSA/ICTP in Trieste).
Twitter @denocris, GitHub.
I am interested also in non-language/exotic applications of NLP (genomics), Quantum Machine Learning, and whatever technology has a positive impact on one of the United Nations Sustainable Development Goals.
I am open to discuss and share our experience in NLP. I am also open to taking part in cool open source projects!
See you soon!
I am Marco Postiglione, a first year PhD student @ UNINA (Naples).
My PhD course is focused on the application of Information and Communication technologies to the healthcare domain. In particular, my project aims to extract valuable information from the unstructured text of Italian clinical notes.
I am also interested in graph analytics (my first publication is on the application of GCNs and LSTMs to Italian mobility data at the beginning of the pandemic) and eXplainable AI.
Anche io carichissimo per il percorso nell’NLP iniziato quest’anno e per la nascita di questa community!!
I am pleased to see that somebody decided to open this thread for the whole Italian community!
Let me introduce myself:
My name is Alberto Bellini, and I am one of the founders of Neuraly AI, a fast-growing startup committed to designing AI-Driven solutions and services through the most advanced Machine Learning and Data Science technologies.
Our R&D team is constantly seeking new challenges, and we actively contribute to the Huggingface community (and even Explosion, creators of SpaCy). For instance, our Italian Sentiment Analysis Model is the first of its kind for the Italian language, so make sure to be among the first ones to try it out!
NLP is one of our core businesses, and we are specialized in exploiting any kind of Transformer architecture to solve several different problems that affect our daily lives. We are one of the first Italian startups that tries to accelerate digital innovation in Italy through SotA NLP & AI solutions!
For the most curious out there: I studied at Politecnico di Milano and the University of Illinois of Chicago, where I took my B.Sc and two M.Scs in Computer Science and Engineering, specializing myself in NLP & Computer Vision.
I do look forward to hearing from the Italian community!
GitHub => @neuraly-ai / @ab3llini
Linkedin => Neuraly
Salutamu! Sugnu cuntentu di essiri ccà, puru ca nun capisciu lu talianu. Scusatimi! Studiava a Bologna, ma 'n goliardia. Eccu picchi li mè studi universitari mi lassaru cuntentu, ma analfabettu.
Staiu sviluppannu un dizziunariu e un tradutturi pi la lingua siciliana. E mi piacissi assai sviluppari quacchi cosa simili pi l’autri lingui taliani.
Si la lingua è la chiavi ô cori umanu, amâ rapiri quacchi cori! Cuminciamu cu puisia. Addistrannu un modellu a truvari versi tradotti, mustramu a li taliani la biddizza dî sò lingui.
Chi ni pinzati? Cu boni dati, ottinemu boni risultati. E cu bedda puisia, ottinemu risultati puetici.
Sugnu cuntentu di scriviri puisia cu vuiatri!
I’m Daniele Pace, currently working as AI Engineer, mostly on NLP tasks such as Text Classification and Data Extraction with bank-related documents (in Italian). Here my website with a short collection of open-source projects.
I attended as Fellow Engineer the last session at Pi School of AI, where I had the chance to meet, directly or indirectly, a lot of extraordinary researchers and engineers (few of them in this post ), and I’m thrilled about building an NLP community in Italy.
By my own, I’m developing a model to generate text, given an input, on the style of Dante Alighieri using GroNLP/gpt2-small-italian · Hugging Face ; it’s almost done, I’m just refining it using some techniques to improve syntax coherence. In short I’ll upload it in the Space Hub of HF and everyone can play with it .
I’m passionate about the intersection between AI and Linguistics, and currently I’m reading “La ricerca della lingua perfetta” by Umberto Eco: it is awesome and definitely an enlightening book for every NLP engineer.
See you shortly!!
Salve, avrei bisogno di effettuare il training di gpt-2 su un corpus della lingua italiana. Qualcuno mi può spiegare dove trovare indicazioni? Non mi interessa se i risultati saranno scarsi. Ringrazio per qualsiasi indicazione