Team of Romanian researchers develop open source language model dedicated to Romanian language

Autor: Cătălin Lupășteanu

Publicat: 14-05-2024

Actualizat: 14-05-2024

Article thumbnail

Sursă foto: stiripesurse.ro

A team of researchers from the Politechnica University of Bucharest and the Institute for Logic and Data Science has developed a language model dedicated to the Romanian language, which can be used for the development of tools and platforms with Artificial Intelligence (AI), was developed by a team of Romanian researchers from Bucharest Polytechnic, University of Bucharest and the Institute for Logic and Data Science.

According to a press release of the BRD Groupe Societe Generale, the banking institution that supported the project, the model is "open source" and can be accessed and used by anyone who wants to build tools based on artificial intelligence.

At the same time, with the publication of this LLM (Large Language Model), the initiators of the project also launch the OpenLLM-Ro community, which aims to together all those who want to contribute in various ways to the development of AI technologies for the Romanian language.

Both projects are initiated and run by the Politehnica University, the University of Bucharest and the Institute for Logic and Data Science, with the support of BRD Groupe Societe Generale.

"The Romanian model launched today is the adaptation of a public LLM developed mainly for the English language. But it was exposed to several million documents in the Romanian language, in order to better understand the meaning of the words. This is essential for the performance of such models in situations where the expression of the request or question entered by the user, as well as the answer, must be in Romanian. Starting from the second part of 2023, a team of researchers from the Politechnica University of Bucharest worked on the development and training of this LLM, and the University of Bucharest and the Institute for Logic and Data Science contributed with researchers who worked pro-bono and, in addition, the Politechnica University provided the computing power needed to train the model. The main partner of the project is BRD Groupe Societe Generale, which supports innovation and the technologies of the future in Romania in all their forms," the press release states.

The effort to specialize a model dedicated to a language is frequently coordinated by the academic community associated with that language, recent examples being from countries such as France, Germany, Spain, Finland, Bulgaria.

The Romanian language model represents the adaptation of a public LLM developed mainly for the English language, but which was exposed to several million documents in the Romanian language, in order to better understand the meaning of the words.

Google News
Comentează
Articole Similare
Parteneri