Artificial intelligence research report

How far has artificial intelligence research developed over the past two decades? Global information analysis company Elsevier shares its findings.

The report's data shows how China will become a global leader in artificial intelligence research, 2004 has surpassed the US in the year and quickly shut down Europe's leading position in this field. At this rate, the report predicts that the country will become the largest source of global artificial intelligence research within four years.

The main findings of the report include:

  • Europe is suffering from AI Brain Drain: during this 20 year, Europe has been losing academic talent in this area, mainly in the corporate sector in the United States.
  • Globally, research on artificial intelligence has increased 5% annually in the past 12.9 years.
  • 20 has had several key milestones over the years that have led to a surge in artificial intelligence research and growth.
  • The ethics of artificial intelligence is a blind spot in research. Although the social relevance of artificial intelligence is getting higher and higher, and the media is paying more and more attention to moral influence, research is still limited.

The report shows that on a global scale, artificial intelligence research has grown by more than 2013% per year in the past five years (2017-12), compared with less than 2008% in the previous five years (2012-5). In contrast, in the past five years (2013-2017), the total global research output value in all disciplines has increased by 0.8% annually. Enrico Motta, Professor of Knowledge Technology at the Open University of the United Kingdom, is one of the many experts who contributed to the report. He believes that this is the most comprehensive feature of AI output in different fields delivered so far: "The report conducted extensive text mining and semantic analysis on documents from different fields to reveal how to more fully define the AI ​​field – AI is mainly used to map AI," he added.

Reviewing the 600 documents of four departments (research, education, technology and media) and 700 keywords in specific areas, semantic analysis shows that the field of artificial intelligence focuses on seven different research areas:

  • Search and optimization,
  • Fuzzy system,
  • Natural language processing and knowledge representation,
  • Computer vision,
  • Machine learning and probabilistic reasoning,
  • Planning and decision making, and
  • Neural Networks.

In these areas, machine learning and probabilistic reasoning, neural networks and computer vision studies have shown the largest research output and growth.

Keyword clustering and co-occurrence in the field of artificial intelligence
Keyword clustering and co-occurrence in the field of artificial intelligence in 2017; Source: Scopus

The report highlights the results of the regional survey:

  • In terms of research areas in the field of artificial intelligence, Europe is the largest and most diverse region with a high level of international cooperation.
  • After China and the United States, India became the third largest producer of artificial intelligence research, followed by Germany and Japan.
  • Iran ranks ninth in the production of artificial intelligence publications, comparable to France and Canada.
  • The international mobility and cooperation model shows that China is relatively isolated from the broader research community.

“By comprehensive research on the performance of artificial intelligence research, our goal is to provide insights into the dynamics, trends and parameters of the field,” said Dan Olley, CTO of Elsevier. “The report is not a conclusion, but about how we best get into manual work. The beginning of the discussion of the intelligent age and increasingly symbiotic technology," he concluded.

The data used in the report comes from Elsevier's Scopus, Fingerprint Engine, PlumX, ScienceDirect and SciVal, RELX's TotalPatent, and further utilization of public resources, including dblp, arXiv, Stanford AI Index, and Kaggle, and datasets by the Chinese Academy of Sciences Provided by the Institute of Automation.

This article is transferred from medium,Original address