🎬Gure zentroko zuzendaria, @eagirre, Jakiunden izan da, Txema Pitarkerekin batera egindako solasaldi batean. Adimen artifiziala, ChatGPT, euskararen erabilera ikerketan eta zientziaren erantzukizuna izan dituzte hizpide.
🎥 Solasaldia osorik 👉https://t.co/rZzayZCUYG
@jakiunde
#Latxa hizkuntza-eredu eta txatbota zer den ulertzeko elkarrizketa grabatu dugu, Latxa bera egin dutenen eskutik.
Ez baduzu probatu Latxa, hemen duzu: https://t.co/GD7JHHjJ0g
This week we held the final meeting of the TRAIN project and the kick-off meeting of the AI4I project. It was a pleasure to work with the @atlanTTic_uvigo and @BSC_CNS teams, and we hope the new collaboration will be just as fruitful. #proyectoTRAIN
We are happy to announce the next webinar in the Language Technology webinar series!
Speaker: Jose Camacho-Collados (Cardiff University)
Title: How Language Models Navigate Culture in a Multilingual World
Date: Thursday, October 16, 2025 - 15:00 CET
Euskara teknologizatzen lagundu nahi? Sartu https://t.co/aamxCMJf8Y gunera eta esaiguzu zer deritzezun euskarazko txioen itzulpenari. Animatu! Zure laguntza behar dugu!
We are happy to announce the next webinar in the Language Technology webinar series organized by the HiTZ Chair of AI&LT!
Speaker: Sampo Pyysalo (University of Turku)
Title: Towards Open Foundation Models for Europe
📅Tuesday, May 13, 2025 - 10:00 CET
🧙♂️ New paper 🧙♀️:
Presenting Wicked: a simple automated method to make MCQA benchmarks more challenging. Wicked shook up 18 open-weight LLMs on 6 benchmarks, with up to 19.7% performance drop with direct prompting 🤯
Paper: https://t.co/KR42BEs0As
Code: https://t.co/94BhRPqCH1
Euskarazko txatbot publikoak hobetzeko zure laguntza behar dugu! https://t.co/lcYpOMCK91 Sartu, galdetu eta erantzun hoberena aukeratu. Erraza eta dibertigarria, lehiaketa eta sariak daude!
.@huggingface-k antolatzen duen nazioarteko ekimen honetan, adimen artifizial irekia munduko edozein hizkuntzara hedatzea du helburu.
#Euskara barne!
Lehen ataza testu batzuen kalitatea neurtzea da, zehazki, hezkuntzarako baliagarriak diren edo ez esatea. Adibideak hari honetan
Euskaraz dakiten adimen artifizial irekiak hobetzen lagundu nahiko zenuke?Zure laguntza behar dugu!Ez da esperientzarik behar - besterik gabe, jarraitu esteka, kontua sortu eta hasi anotatzen!
https://t.co/r7WD5E2dQb
Zein da hizkuntza eredurik onena euskaraz? Zer moduz ote dabil gure Latxa 🐑 txiki berria erraldoi komertzialekin konparatuz? Galdera hauei erantzuteko esperimentu bat prestatu genuen atzo, eta emaitzak oso onak izan dira. Eskerrik asko parte hartu duzuen guztioi! #IEB2024
We are launching Salamandra 2B & 7B multilingual LLMs trained at @BSC_CNS from scratch with nearly 8 trillion tokens in 35 EU languages+code. Spanish languages have been carefully curated, with Romance languages comprising >30% of the training dataset.
https://t.co/g6WEWvVmSs
👇
Ba al dakizu Xuxenek 30 urte bete dituela? Bai, bai! 🎉🎉🎉
Xuxenen historiari buruz gehiago jakin nahi baduzu, @euwikipedia-n artikulu ederra daukazu
https://t.co/e8aIZq58vF
Eta gehiago prentsan hurrengo orduetan!
Excited to be giving this course on the foundations of deep learning for NLP, which will be later extended to Large Language Models in two further courses, and complemented with an application oriented course. Learn from top researchers who actually build, use and research on these models. With @I_MirandaM@4nderB@oierldl and @ragerri among others.
New @hitz_zentroa paper! "MedExpQA: Multilingual Benchmarking of Large Language Models for Medical Question Answering"; w/ @ragerri and @jiporanm in Artificial Intelligence in Medicine (Elsevier)
Paper: https://t.co/iXH8ovdamD
Resources: https://t.co/pcBAF2FKfN
Happy to see that the SignON project continues to receive recognition. Looking forward to continuing to collaborate with the fantastic consortium on future projects.
SignON project, led by Prof. Andy Way (@tarfandy), wins Most Innovative Transformation award at Digital Transformation Awards 2024!🙌They developed a communication service using machine translation to bridge spoken & sign languages.
Read more➡️ https://t.co/VovxlMTmlE