Japan accelerates development of “Made in Japan” AI model

The Japanese government and SoftBank are joining forces to create an AI model trained entirely on domestic data, aiming for technological sovereignty and national security.

Japan is aiming to build a domestic AI model to reduce its dependence on technology from the US and China.

SoftBank, through its subsidiary SB Intuitions, is developing the Japanese model “Sarashina” using more than 10,000 GPUs on Nvidia’s DGX SuperPOD platform.

The project aims to train AI using local data that accurately reflects Japanese context, culture, and language.

The Tokyo government has pledged strong support as this is not just a matter of technology but also of protecting national data sovereignty .

Corporations like Dentsu and SoftBank have joined forces to develop LLMs that create Japanese-language advertising content, leveraging domestic creative data.

However, Japan faces big problems in terms of cost, training scale and lack of high-quality AI human resources.

Instead of pursuing giant models like GPT-4, Japan chose the direction of “big enough to be useful, small enough to be effective”.

If successful, the “Made in Japan” AI model will help Tokyo lead the trend of localized AI, where technology speaks the language and culture of its users.

Dear readers, please watch more videos : AI Trash Cleaning | Hanoi 18:00

Source: https://khoahocdoisong.vn/nhat-ban-tang-toc-phat-trien-mo-hinh-ai-made-in-japan-post2149062978.html