Pietro Michiardi retweeted
What a day at the #AI #Summit! Thrilled to report that @EURECOM won the AI Frugal Challenge organized by @huggingface !
Our task: detect misinformation about climate change in texts
The catch: be accurate but be very energy savvy!
More to come with #ClimateSense, a new EU project
Pietro Michiardi retweeted
Happy to share that our paper, "Information Theoretic Text-to-Image Alignment," has been accepted at ICLR 2025! 🚀
Kudos to the team: Chao Wang, the PhD student leading this work, @FranzeseGiulio , @michiard and @_gallomax !
arxiv.org/abs/2405.20759
Pietro Michiardi retweeted
Meet EULER (@EURECOM Language Model for Education and Research) and our new #Socratic LLM for AI-Driven Education, eurecom-blog.medium.com/eure…
Model available @huggingface and @ollama
#LLM #RAG #EduTech
Pietro Michiardi retweeted
It is an absolute honor to have DNA generated by our work at @EURECOM exhibited at Velarde as a part of DNA:AND exhibit of Malcolm Le Grice--the most influential modernist filmmaker in British cinema. velarde.co.uk/exhibitions/19…
Pietro Michiardi retweeted
"Fine Tune a Large Language Model for Socratic Interactions" is presented at #AI4EDU workshop at @kdd_news #KDD2024
Read giovannigatti.github.io/socr…
Code: github.com/GiovanniGatti/soc…
Model: huggingface.co/eurecom-ds/Ph…
EULER: euler.eurecom.fr/
with @paolopapotti @michiard at @EURECOM
Pietro Michiardi retweeted
Presenting our work on tabular time-series today @kdd_news! Drop me a DM if you're interested to chat
Thanks a lot to Zied, @_gallomax, @finale80 and @michiard
Paper and code: arxiv.org/abs/2406.15327
Pietro Michiardi retweeted
🌟 Exciting News! Our paper "SΩI: Score-based O-INFORMATION Estimation" is accepted for an oral presentation (Top 1.5% ) at #ICML2024 ! 🎉. Join us on July 25 for the oral (11:00-11:15 AM, Hall A2) & poster session (11:30 AM-1:00 PM, Hall C 4-9 1702).
proceedings.mlr.press/v235/b…
Pietro Michiardi retweeted
Thrilled to announce that our paper "SOI: Score-based O-INFORMATION Estimation" has been accepted to #ICML2024 in Vienna! 🎉📚
arxiv.org/pdf/2402.05667
TL;DR: A new method to estimate O-information using score functions to describe high interdependencies in complex systems.
Thrilled to announce that our paper "SOI: Score-based O-INFORMATION Estimation" has been accepted to ICML 2024 in Vienna! 🎉📚 arxiv.org/pdf/2402.05667
- “Masked Multi-time Diffusion for Multi-modal Generative Modeling" in NeurIPS's 2023 Workshop on Diffusion Models. eurecom.fr/publication/7540/…
Pietro Michiardi retweeted
Full room for @rtroncy talk on CIMPLE: Hybridizing #KnowledgeGraph and #LLM for Explaining Factors Contributing to Misinformation in Montefiore institute in Liège. The talk is part of our KG4DI research network in Belgium. w3id.org/kg4di @chrdebru @Ben__DM @CoenDeRoover
Pietro Michiardi retweeted
On booth C13 at #waicf2024, discover EULER, the @EURECOM Large Language Model for Education and Research. Our LLM adopts à Socratic approach to guide the students to find answers to their questions. The LLM is robust generating its answers from our course slides using RAG
Pietro Michiardi retweeted
Replying to @EURECOM
@EURECOM will be at #NeurIPS2023 in full force. On my way to present our work on Qatch (@NeurIPSConf) and Galois (@TrlWorkshop). Please reach out if you'd like to meet
Pietro Michiardi retweeted
in his AI-ready Healthcare podcast:If you are interested in #Trustworthy #MedicalAI, take some time to listen to the fun chat I had with@anirbanakash in his Ai-ready Healthcare podcast:
podcasters.spotify.com/pod/s…
Pietro Michiardi retweeted
As we progress in @syclopseu, we are setting up a technical blog to share details about all the cool #RISCV and #SYCL work that is being done in the project. As a teaser, we open the blog with a quick partner introduction. Watch this space for more updates soon!
The SYCLOPS project is well underway, and our partners are working hard on cutting-edge research that brings together the open #RISCV and #SYCL standards.
We've introduced the work of each partner in this blog!
syclops.org/updates/2023/11/…
Pietro Michiardi retweeted
We wouldn’t be computer scientists if we also didn’t find a way to digitally encode smells and capture them in DNA thanks to a collab with the @oligoarchive #H2020 #ProjectCrossover #SmellCultureFair
This is indeed a very nice paper, and the code works out-of-the-box! We used it in our work (MINDE: Mutual Information Neural Diffusion Estimation arxiv.org/abs/2310.09031) , and we were absolutely happy to have this benchmark!
Thank you for your work and see you at #NeurIPS2023!
Are you developing a new #MutualInformation estimator, or want to know which type of estimator is preferable for certain distributions?
🧐Then check out our #NeurIPS2023 paper "Beyond Normal: On the Evaluation of Mutual Information Estimators"!
📄 openreview.net/forum?id=25vR… 🧵1/4
























