Beijing AI academy unveils world's largest pre-trained deep learning model
Published: Jun 02, 2021 02:49 PM
AI Photo: CFP

AI Photo: CFP

The Beijing Academy of Artificial Intelligence (BAAI) unveiled a newer version of its hyper-scale pre-trained deep learning model, the country's first and the world's largest, at an ongoing AI-themed forum in Beijing, in the latest signal of China's ambition to become a global leader in AI.

The latest version of the model, known as Wudao, literally meaning an understanding of natural laws, sports 1.75 trillion parameters, breaking the record of 1.6 trillion previously set by Google's Switch Transformer AI language model, the academy announced Tuesday at the three-day forum that runs through Thursday. Wudao was only initially released in March.

Wudao is intended to create cognitive intelligence dually driven by data and knowledge, making machines think like humans and enabling machine cognitive abilities to pass the Turing test, Tang Jie, BAAI's vice director of academics, said during the forum.

The newer version of Wudao is both gigantic and smart, featuring its hyper scale, high precision and efficiency. A deep learning system called FastMoE, developed by BAAI researchers, is disclosed to have laid the cornerstone for Wudao to break through foreign technological bottlenecks and launch the trillion-parameter model.

Huazhibing, a virtual student developed based on Wudao 2.0, the first of its kind in the country, also debuted at the forum on Tuesday, with a video revealing her showcasing music, poem and painting for the forum.

Currently, the combination of a large-scale model and immense computing power offers a plausible road map to achieving artificial general intelligence, said Zhang Hongjiang, chairman of BAAI, envisioning the availability of large-scale model-based transformative AI industry infrastructure resembling power grids in the future. 

A large-scale AI model is analogous to a power plant, which is supposed to convert data, or fuels, into smart capacities to power various AI applications, Zhang explained.

If the large-scale model connects with all AI applications to offer users unified smart capacities, the whole society will be equipped with an "intelligent network," according to the former chief executive of Chinese software major Kingsoft, describing such models as strategic infrastructure underpinning AI development.

Global Times