The Institute of Automation of the Chinese Academy of Sciences released the new generation of artificial intelligence large model developed by the institute in Shanghai this morning (16th) - Zidong Taichu 2.0. Compared with the first generation, the brand-new large model focuses on improving the decision-making and judgment capabilities, realizing the leap from perception, cognition to decision-making, and will play a greater role in the fields of medical care, transportation, and industrial production in the future.

The artificial intelligence large-scale model is a huge and complex network system, similar to the human brain. The more nodes and data in it, the stronger the processing power. Usually, the number of nodes in this large-scale model is more than tens of billions. As my country's first graphic, text, and audio three-modal large-scale model that stores hundreds of billions of parameters, the first-generation Zidong Taichu AI large-scale model realized the ability to recognize images, text, and voice without logos when it came out in 2021. The new generation of large models has been upgraded and integrated into recognition capabilities such as video, sensor signals, and 3D point clouds, breaking through key technologies such as multi-modal group cognitive coding, decoding, and full-modal cognition, and improving system decision-making and judgment. Ability, which opens up the path of system recognition ability from perception, cognition to decision-making, thus expanding a wider range of application scenarios. At present, the large model of Zidong Taichu has been applied in many fields such as neurosurgery navigation, short video content review, medical multi-modal differential diagnosis, and image study of traffic violations.