The Paper Journalist Zhang Wei
The latest report shows that China has ranked first in the world in the number of artificial intelligence patent applications in the past decade, and deep neural networks, feature extraction, and image classification are hot research areas.
On January 20th, Tsinghua University Artificial Intelligence Research Institute and Tsinghua-Chinese Academy of Engineering Knowledge Intelligence Joint Research Center jointly released the "Artificial Intelligence Development Report 2020".
The report predicts that the key development directions of artificial intelligence in the next decade include reinforcement learning, neuromorphic hardware, knowledge graphs, intelligent robots, interpretable AI, digital ethics, and knowledge-guided natural language processing.
The "Artificial Intelligence Development Report 2020" is based on the "Science and Technology Intelligence Big Data Mining and Service Platform" (AMiner) platform independently developed by Tsinghua University. It contains papers and patent data from top journals and conferences in the field of artificial intelligence from 2011 to 2020. , Trying to show the major scientific research progress, achievement output and scientific research hotspots obtained so far in the development of artificial intelligence.
According to the papers published in the top international journals and conferences in the field of artificial intelligence in the past 10 years, the report selects the top ten AI research hotspots in the past ten years, including deep neural networks, feature extraction, image classification, target detection, semantic segmentation, representation learning, Generate confrontation networks, semantic networks, collaborative filtering and machine translation.
Top 10 AI research hotspots in the past decade
The report also summarized the research status and trends in key areas of artificial intelligence such as machine learning, natural language processing, and knowledge engineering, and pointed out that pre-trained language models have made important progress in the field of natural language processing. "Since the emergence of a series of pre-trained language representation models such as ELMo, GPT, BERT, etc., pre-training models have shown far more effects than traditional models on most natural language processing tasks, and have received more and more attention. It is one of the biggest breakthroughs in the field of natural language processing in recent years."
Development Trends of Technology Research in the Field of Natural Language Processing
The report found that China's scientific research output level in more than 10 AI sub-fields such as natural language processing, chip technology, machine learning, information retrieval and mining is among the highest in the world; the output of papers in the field of multimedia and Internet of Things exceeds that of the United States, Number one in the world; and in the fields of human-computer interaction, knowledge engineering, robotics, computer graphics, and computational theory, China still needs to catch up.
The report conducted an analysis of high-level talents in the global AI field in Chapter 6, and pointed out that the United States has the largest number of high-level AI scholars, with 1244 person-times, accounting for 62.2%, which is the second country (China) in the number of high-level AI scholars 6 Times more. The number of high-level scholars in China is 196. Germany ranks third, the country with the largest number of scholars in Europe, and the number of scholars in other countries is less than 100. The report also pointed out that high-level talents in the field of artificial intelligence in China are mainly distributed in the Beijing-Tianjin-Hebei, Yangtze River Delta and Pearl River Delta regions.
Top 10 countries for the number of high-level scholars in the field of artificial intelligence
According to the report, China has begun to focus on the development of high-level talents in the AI field since 2018, mainly by colleges and universities through the establishment of AI academies and AI majors.
Up to now, a total of 215 universities and colleges in China have established undergraduate programs in "artificial intelligence". Among these universities, there are 60 double-class universities (28%), and the other 155 are ordinary undergraduate universities. As of June 2019, at least 38 colleges and universities have established independent artificial intelligence colleges to fully carry out undergraduate and postgraduate education, and begin to recruit undergraduates in artificial intelligence in 2019.
Artificial intelligence patent applications are also one of the focus of this report. According to statistics, the number of global artificial intelligence patent applications in the past decade was 521,264. The number of patent applications in China was 389,571, ranking first in the world, accounting for 74.7% of the global total, 8.2 times that of the second-ranked US patent applications.
Top 10 countries in global AI patent applications
Based on the analysis of the maturity curve of artificial intelligence technology and the development status of artificial intelligence in 2020, the report predicts that the key development directions of artificial intelligence in the next ten years include reinforcement learning, neuromorphic hardware, knowledge graphs, intelligent robots, interpretable AI, and digital Natural language processing guided by ethics and knowledge, etc.