PulseAugur
EN
LIVE 23:45:29

New multilingual text-to-image AI model NeoBabel outperforms Big Tech alternatives

Dheeraj Varghese, a researcher at the Universiteit van Amsterdam, has developed NeoBabel, a novel multilingual text-to-image model. This model operates directly in six languages, bypassing the need for English translations, and has demonstrated superior performance compared to existing Big Tech alternatives. Varghese utilized the LUMI supercomputer, accessed via SURF, to train NeoBabel on an extensive dataset of 124 million image-label pairs, aiming to enhance AI inclusivity. AI

IMPACT This model could significantly improve AI accessibility and inclusivity by enabling direct interaction in multiple languages, reducing reliance on English as a universal intermediary.

RANK_REASON Development of a new AI model with novel capabilities. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — sigmoid.social →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

New multilingual text-to-image AI model NeoBabel outperforms Big Tech alternatives

COVERAGE [1]

  1. Mastodon — sigmoid.social TIER_1 English(EN) · [email protected] ·

    What if AI finally spoke your language? Dheeraj Varghese from the Universiteit van Amsterdam developed NeoBabel a multilingual text to image model that works di

    What if AI finally spoke your language? Dheeraj Varghese from the Universiteit van Amsterdam developed NeoBabel a multilingual text to image model that works directly in six languages without English translations. With the computing power of the European supercomputer LUMI access…