Talent drain: how Xavier Niel wants to bring French AI geniuses back into the fold


Event ai-PULSE (Paris) – “I hope that tomorrow our children will use algorithms developed in France and equipped with our cultural specificitiess”. Noting that the best French researchers in artificial intelligence have left to join the private laboratories of GAFAM, Xavier Niel asked himself on stage the question of how to bring them back to the country.

Thus, the founder of Free does not intend to let the United States and China, ahead in the field of AI, impose their vision of the world. “ I would like us to talk about French imperialism in AI in the future.”

To make this dream come true, the answer is called Kyutai (pronounced “Cute AI” and meaning “sphere” in Japanese), a non-profit laboratory entirely dedicated to open research in artificial intelligence. By making the products of its fundamental research public, it aims to contribute to the democratization of AI. This so-called open science mode is opposed to the closed mode of operation of Meta or Google DeepMind laboratories.

Xavier Niel intends to use this as an argument to bring French geniuses back across the Atlantic, beyond the quality of life in Paris. With nearly 300 million euros, Kyutai also has the means to achieve its ambitions. Its three co-founders, the iliad group (Free), the CMA CGM group and Schmidt Futures, the philanthropic company of Eric Schmidt, former CEO of Google, have each put in the pot.

“There are a lot of ideas in France but not enough moneydeplores Rodolphe Saadé, director of the shipowner CMA CGM. With this envelope, unique in Europe, we fill this gap. » The majority of this funding will go towards purchasing the computing power needed to run future models. Other private investors are now invited and other entities to join them to finance the work of this “non-profit” organization in the long term.

Yann Le Cun for godfather

Xavier Niel will serve as the first president of Kyutai, whose head office will be based in the third arrondissement of Paris. To guide its work, the laboratory has a scientific council which includes Yann Le Cun, founder of Facebook AI Research (FAIR). He will then share the results of his research with the scientific community, the startup ecosystem and more broadly to all enlightened citizens. Kyutai will also train future experts in the discipline, by hosting master’s student internships and supervising doctoral and post-doctoral students.

Open science consists not only of sharing the source code of a model but also the code that was used to train it. “ It is often this “secret sauce” that explains the success of the model », explains Hervé Jégou. This former INRIA and Facebook employee is one of the first six French researchers to join the Kyutai adventure with Alexandre Défossez (FAIR), Edouard Grave (Apple MLR, FAIR), Laurent Mazaré (DeepMind), Patrick Pérez (Microsoft Research, Valeo) and Neil Zeghidour (DeepMind).

Create a new ChatGPT “from scratch”

What work will they work on in the coming months? Unlike Google DeepMind which tackles very concrete problems such as weather forecasting or molecular biology, Kyutai will remain in the field of fundamental research.

One of the avenues mentioned concerns the development of large multimodal models which generate text and images but also sound, by providing them with visual and auditory perception. An AI will learn, for example, to draw by giving feedback to the apprentice designer on the current sketch.

Kyutai intends to create, “from scratch”, a large language model (LLM), such as ChatGPT from OpenAI or Bard from Google in order to master the technical base and intervene in all stages of design and training. It is also about offering an alternative to the “Transformers” models used in most computer vision or text generation applications. Finally, effort will be made on the explainability of the models to avoid the “black box” effect.



Source link -97