How to be as 'smart' as Auto-Regressive LLMs:
- memorize lots of problem statements together with recipes on how to solve them.
- to solve a new problem, retrieve the recipe whose problem statement superficially matches the new problem.
- apply the recipe blindly and declare…

Cinquenta Anos Após o 25 de ABRIL: Celebração Prospetiva

Com transmissão em direto no YouTube da Academia das Ciências de Lisboa

👉 Inscreva-se através do seguinte link: docs.google.com/forms/d/e/1FAI…

#25deAbril #CiênciaPortuguesa #AcademiaCiênciasLisboa

Looking fwd to seeing you in Lisbon, Mohit Bansal!

Reminder that the LxMLS application deadline is 3 days away! lxmls.it.pt/2024

This year with Kyunghyun Cho Albert Gu Mohit Bansal Noah A. Smith Sara Hooker Danqi Chen Desmond Elliott and other amazing speakers.

📢 Only 10 days left to apply to the Lisbon Machine Learning School in July. Check out our fantastic lineup of speakers 👇

🔗 for more information: lxmls.it.pt/2024/

🔗 to apply: tinyurl.com/288raakv

#LxMLS #MachineLearning #NLP

Day #4 of #ALPS24 : nature activities (again), great talk from ⁦Andre Martins⁩ on machine translation (and its evaluation) using LLMs, a lab on RAG by ⁦Matthias Gallé⁩, an evening poster session and we finished by a movie night 🍿! What a day !

4th Advanced Language Processing Winter School #ALPS24 starts Mon in breathtaking Vanoise (FR alps)😏! Fab speakers Sara Hooker Barbara Plank Thomas Wolf Andre Martins Claire Gardent Maxime Peyrard & jolly organisers LIG NAVER LABS Europe cohere Info: lig-alps.imag.fr

Wonderful talk this morning by Chryssa Zerva at #UncertaiNLP at #EACL2024 about the different aspects of uncertainty in #NLProc and her work on conformal prediction!

Are you attending the eaclmeeting? Don't miss Chryssa Zerva's keynote talk on 'Uncertainty in NLP: Quantification, interpretation and evaluation' at the UncertaiNLP workshop on the 22nd March.

🔗 more info uncertainlp.github.io/program

Attending #EACL2024 ? Check out our paper with Dennis Ulmer (is on the job market 👨🏻‍💻) and Chryssa Zerva on non-exchangeable conformal prediction for generation.

An LLM for law by Pierre Colombo Telmo and the Equall team!

Super proud about this work! Amazing collaboration between Unbabel Técnico Lisboa CentraleSupélec to take multilingual LLMs and MT to the next level!

Today we release the Tower paper! 🗼
Tower is an open-weight suite of multilingual models — built on top of LLaMA-2 — for translation-related tasks. It supports 10 different languages.

Paper: arxiv.org/pdf/2402.17733…
Models and data: huggingface.co/collections/Un…
🧵Thread below.

I have signed this document, prepared under the leadership of Monica Bettencourt-Dias Joana G Sá and Isabel Rocha (UNL) 👇

Compromisso de Regime para a Ciência e Inovação 2024-2034 publico.pt/2024/02/27/cie…

Fenchel-Young losses meet Hopfield nets -- very excited about this work led by Saul Santos.

Super-excited about our Tower models which keep getting better and better!

Conformal prediction meets kNN for language generation! To be presented at #EACL2024 .

Interview about the ELLIS program in NLP which Iryna Gurevych Ivan Titov and I are co-directing.