Recently, the 2021 Global Artificial Intelligence Technology Conference was held in Hangzhou, China. At the Computer Vision Forum, Huang Feiyue, Deputy General Manager of Tencent Youtu Lab, officially announced that Tencent Youtu Lab and the Institute of Artificial Intelligence, Xiamen University released the "Top 10 Artificial Intelligence Trends in 2021" (hereinafter referred to as the Trends Report). Based on their long-term research observation on artificial intelligence, especially computer vision, the Trends Report puts forward cutting-edge predictions in several fields such as 3D vision technology, digital content industry, AI deep learning algorithms and AI core chips.

The Trends Report points out that with the further improvement of automation and interpretability of automatic machine learning and the increasing perfection of the entire architecture of AutoML, the construction of a new generation of universal AutoML platform will be promoted, ultimately realizing the popularization of machine learning. In the past few years, the great success of deep learning cannot be achieved without large-scale annotated data sets. However, with the continuous expansion of business scale, the cost of data annotation has become one of the main factors impeding enterprises' cost reduction. Besides, unsupervised and weakly supervised learning will become a new tool adopted by enterprises to address this obstacle.

At the conference, Huang Feiyue said that with the continuous upgrading of algorithms and hardware capabilities, 3D visual intelligence technology will facilitate the development of business and individual consumption. Deep learning is moving towards multimodal fusion, edge computing is integrating with artificial intelligence expeditiously, and the equity research of AI algorithm will drive the application of AI towards universal benefit.