The specialists of the Academy of Artificial Intelligence of Beijing (China) this week presented the most sophisticated natural language processing model in the world, which uses 1.75 trillion parameters to simulate conversational speech, write poems, understand images and even generate recipes, collects the South China Morning Post newspaper.
WuDao 2.0, which means 'understanding of natural laws' in Chinese, is a previously trained artificial intelligence model that was developed with the help of more than 100 scientists. It is more powerful than the models of its main competitors: the GPT-3 from the company OpenAI (co-founded by Elon Musk), which was launched with 175,000 million parameters, and Google's Switch Transformer, which uses 1.6 trillion parameters.
The model develops acquired skills in both Chinese and English, as it has studied 4.9 terabytes of images and texts, including 1.2 terabytes of texts in those two languages. WuDao 2.0 already has 22 partners, such as smartphone maker Xiaomi or short video giant Kuaishou.
The path to general artificial intelligence is large models and large computers, said the president of the Beijing Academy of Artificial Intelligence, Zhang Hongjiang.
What we are building is a power plant for the future of artificial intelligence, with mega data, mega computing power, and mega models; we can transform the data to power the artificial intelligence applications of the future, Hongjiang said.
The presentation of WuDao 2.0, the first version of which was released last March, comes as China and the US are in a race to become the world leader in artificial intelligence and other advanced technologies. Earlier, US President Joe Biden asked Congress to back a $ 13.5 billion increase in total federal spending on research and development.