ArtificiaI Intelligence: How knowledge is created, transferred, and usedTrends in China, Europe, and the United States

Executive Summary

ARTIFICIAL INTELLIGENCE: HOW KNOWLEDGE IS CREATED, TRANSFERRED, AND USED 2

This document summarizes Key Findings from the full report Artificial Intelligence: how knowledge is created, transferred, and used, available alongside other relevant material on the Elsevier Artificial Intelligence resource centre.* The RELX group has extensive data assets, powerful computing capabilities, and a vast technological talent base. These allow Elsevier to provide unique insights on AI through this report. We hope these will be of interest to research evaluators, research funders, policy makers, and researchers, as they seek to navigate this complex, evolving, and fast-growing field.

Artificial Intelligence: a multifaceted field

The increasing importance and relevance of Artificial Intelligence (AI) to humanity is undisputed: AI assistants and recommendations, for instance, are increasingly embedded in our daily lives. However, AI does not seem to have a universally agreed scope. Our bottom-up methodology contributes to the definition and classification of an evolving field with a shifting structure. AI seems to cluster around the areas of Search and Optimization, Fuzzy Systems, Natural Language Processing and Knowledge Representation, Computer Vision, Machine Learning and Probabilistic Reasoning, Planning and Decision Making, and Neural Networks (see Figure 1).

* https://www.elsevier.com/connect/ai-resource-center

3

figure 1 Keyword clusters and co-occurrences in the AI field, 2017; source: Scopus.

Planning and Decision Making

Neural Networks

Natural Language

Processing and Knowledge

Representation Computer Vision

Fuzzy Systems

Search and Optimization Machine

Learning and Probabilistic

Reasoning

ARTIFICIAL INTELLIGENCE: HOW KNOWLEDGE IS CREATED, TRANSFERRED, AND USED 4

While the field spans several domains and can be viewed from different standpoints, such as teaching, research, industry, and media, there seems to be little overlap between these perspectives (see Figure 2). As partly expected, there is less overlap on specialized keywords and more on general terms. Industry tends to emphasize algorithms, possibly for efficient gains in time and human labor. The increasing societal relevance of AI and potential ethical concerns raised by the growing use of algorithms reflect the visibility of applications and ethics themes in the media, which makes AI more imperative and intuitive to the public. Despite this societal relevance, ethics is not yet strongly reflected in the research corpus, although recent conferences reveal a growing focus on ethics. Interestingly, ethics keywords are more heavily represented in teaching, potentially as a result of public interest and some government mandates.

Teaching268

Industry641

Media82

Research42

444

83

52

1533

1

6

1017

figure 2 Keyword mapping (number of keywords) between AI perspectives.

The apparent lack of a common language across perspectives calls into question the quality of understanding and communication across the AI field. With closer and instant collaboration across geographies and sectors, research dialogue shifts away from traditional sequential translation. This results in parallel dialogues, online and through media and social media channels. New stakeholders, such as students, freelancers, and citizens, become involved in research, for example, on competition platforms like Kaggle. A common language and understanding would better connect the AI ecosystem.

AI has emerged as an area of importance for national competitiveness, yet also sees growing international collaboration. Several national and international AI policies and strategies have been put forth in recent years, as both causes and consequences of growing AI research ecosystems (see Figure 3). This has led to increased scientific output through a variety of dissemination modes, including publications, preprints, conferences, competitions, and software. The field has grown annually by 5.3% in the last decade and 12.9% in the last 5 years (see Figure 4).

Global trends in AI research

5

figure 3 Selected AI-relevant policies and events (upper panel) and technology breakthroughs (lower panel), 1998-2018.

201819991998 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

China

Europe

United States

Global

Google Duplex (2018)

First robots for home, e.g. cleaning (2001)

Speech recognition on smartphone (2008)

Google DeepMind winning Go (2016)

First self-driving cars (2005)

Google’s autonomous car (2009)

Evangelist Andrew Ng training an AI (“loving cats”) (2012)

Apples SIRI, Cortana, Google Now (2011-2014)

Europe: new Innovation agenda (EITs) (2014)Launch of Horizon2020 (2014)

Europe: FP7 funding program (2006)

Financial crisis (2008)

Letter against autonomous weapons (2015)

United States National AI R&D Strategic Plan (2016)

President Xi Jinping calling for breakthroughs in S&T (2014)

China National Medium- and Long-Term Plan for the Development of Science and Technology (2004)

1998

Glo

bal n

umbe

r of p

ublic

atio

ns o

n AI

0,000

10,000

20,000

30,000

40,000

60,000

70,000

50,000

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

figure 4 Annual number of AI publications (all document types), 1998-2017; source: Scopus.

ARTIFICIAL INTELLIGENCE: HOW KNOWLEDGE IS CREATED, TRANSFERRED, AND USED 6

There are strong regional differences in AI activity. Europe is still the largest actor in AI research, despite rapid growth and ambition from China, while the US is regaining some ground in recent years (see Figure 5).

figure 5 Share of global publication output in AI (all document types) for periods 1998-2002, 2003-2007, 2008-2012, and 2013-2017, per region; source: Scopus.

0%

20%

40%

60%

80%

100%

Share of world publications in AI

OtherChina Europe United States

1998 – 2002 2003 – 2007 2008 – 2012 2013 – 2017

9%

35%

25%

31%

18%

33%

20%

29%

26%

31%

15%

28%

24%

30%

17%

29%

There are significant differences in citation impact, with the United States leading; however, the three regions have similar download impact (see Figure 6), suggesting comparable usage of each region’s research. While China’s FWCI is still below that of Europe and the United States, it shows tremendous growth over the past two decades, from half the world average to reaching the world average in recent years. Europe’s FWCI remains stable over the period, comfortably higher than the global average. The United States’ FWCI is the highest of the three regions, between one and a half to two times as high as the global average. The 2016-2017 dip in FWCI for the United States may be due to incomplete citation data, although there seems to be a slight decreasing trend following a 2014 peak.

figure 6 Rebased AI Field-Weighted Citation Impact (FWCI, bold lines) and Field-Weighted Download Impact (FWDI, dotted lines) (all document types) per region, 1998-2017; source: Scopus.

0

1.0

1.5

2.5

0.5

2.0

United States

Europe

China

FWCIFWDI

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Regional research trends in AI

7

China aspires to lead globally in AI and is supported by ambitious national policies in the AI arena. A positive migratory inflow of AI researchers in China (see Figure 7) also suggests an attractive research environment. China’s AI focuses on computer vision and does not have a dedicated speech recognition cluster, likely because this type of research in China is conducted by corporations that may not publish as many scientific articles. It shows robust growth of its research and education ecosystems, with a rapid rise in scholarly output and similar research usage as other regions. China’s AI research has a rapidly increasing yet still comparatively low citation impact, which could be a symptom of regional, rather than global, reach. This is also apparent through its relatively low levels of international collaboration and mobility in research (see Figures 7 & 8), which yield a comparatively small but highly cited corpus of AI research. As in many other research areas, collaboration is key to success, as demonstrated by increasing discussions on global social media and growing international AI competition numbers.

figure 7 Share of AI researchers (%) per mobility class, 1998-2017; source: Scopus.

0%

20%

40%

60%

80%

100%

Shar

e of

rese

arch

ers

Sedentary

Migratory Out�ow

Migratory In�ow

Transitory

China

75.7%

17.2%

3.6%3.5%

Europe

52.9%

32.6%

6.8%

7.8%

United States

37.5%

45.0%

8.9%

8.6%

Share of AI researchers per mobility class

ARTIFICIAL INTELLIGENCE: HOW KNOWLEDGE IS CREATED, TRANSFERRED, AND USED 8

figure 8 Number of publications from international collaborations (all document types) and their rebased Field-Weighted Citation Impact, FWCI, 1998-2017; source: Scopus.

0

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.2

0 5% 10% 15% 20% 25% 30% 35% 40% 45%

China

Europe

United States

reba

sed

FWCI

number of publications

figure 9 Academic-corporate share of publications (left-hand side, dark color) and their Field-Weighted Citation Impact, FWCI (right-hand side, light color) (all document types), 1998-2017; source: Scopus.

China

Europe

United States

World3.4%

8.9%

3.6%

2.3%

2.53

3.41

2.46

2.64

Share of academic-corporate publications FWCI of academic-corporate publications

AI knowledge transfer

9

Europe is the largest region in AI scholarly output (see Figure 4), with high and rising levels of international collaborations outside of Europe (see Figure 8), but appears to be losing academic AI talent, especially in recent years. The broad spectrum of AI research in Europe reflects the diversity of European countries, each with their own agenda and specialties. Focus areas of European AI research include genetic programming for pattern recognition, fuzzy logic, and speech and face recognition. Deep learning research in Europe appears less connected to other subfields than it is in other regions, and AI robotics in Europe appear to be embedded in the learning cluster.

The United States corporate sector attracts talent and is strong in AI research and academic-corporate collaboration (see Figure 9), possibly due to their cross-sector joint labs tradition. The United States academic sector is also robust, both in terms of scholarly output and mobility. The country appears to be leading the way in international AI competitions, and United States researchers increasingly collaborate internationally on AI research (see Figure 8). AI in the United States has a strong focus on specific algorithms and separates speech and image recognition into distinct clusters. The corpus shows less diversity in AI research than Europe but more diversity than China.

Among other key contributors in AI, we note the rapid emergence of India, today the third largest country in terms of AI publications after the United States and China (see Figure 10). Iran is ninth in publication output in 2017, on par with countries like France and Canada. Last year, Russia surpassed Singapore and The Netherlands in research output, yet remains behind Turkey. Japan remains the sixth largest in AI research globally, and as the third largest economy in the world, it has an advanced industry sector that is investing in AI applications.

figure 10 Publication output per country/territory (all document types), 2013-2017; source: Scopus.

Number of publications in AI

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

16,000

2013 2014 2015 2016 2017

China United States India United Kingdom

Germany Japan Spain Iran France Italy

ARTIFICIAL INTELLIGENCE: HOW KNOWLEDGE IS CREATED, TRANSFERRED, AND USED 10

In parallel to AI’s growing presence and impact into our daily lives, AI research has pervaded multiple platforms and venues. Beyond traditional journal articles, the report provides insights on other AI channels, such as conferences, pre-print servers, technology competitions, and social media discussions. The seeming underrepresentation of ethics in AI research, despite the urgent imperative for ethical AI, remains one of the most pressing questions posed by the report.

Other questions of interest for potential future investigations are: • Is there a relationship between research performance in

AI and research performance in more traditional fields that support AI (such as computer science, linguistics, mathematics, etc.)?

• How does AI research translate into real-life applications, societal impact, and economic growth?

• Where do internationally mobile AI researchers come from and go to?

• How sustainable is the recent growth in publications?• How will the main players in the field continue to

cooperate and collaborate?• How can we achieve and implement ethical AI?

The full report is available at LINK; for further materials on AI, including the full report from which these key findings are selected, data used in these analyses, interactive graphs, expert opinions, and more, please visit the Elsevier AI Resource Centre (https://www.elsevier.com/connect/ai-resource-center).

11

Alessandro Annoni, Head of Digital Economy Unit, Joint Research Centre, European Commission

Dr. Roberto M. Cesar Jr., Adjunct Coordinator, São Paulo Research Foundation (FAPESP), Brazil

Dr. Yuichiro Anzai, Senior Advisor, Director, Center for Science Information Analysis, Japan Society for the Promotion of Science (JSPS), Chairman, Strategic Council for AI Technology, Japan

Prof. Dame Wendy Hall, Professor of Computer Science in Electronics and Computer Science, University of Southampton, Director of the Web Science Institute, United Kingdom

Prof. Lynda Hardman, Director, Amsterdam Data Science, Past President, Informatics Europe, Research Leader at Centrum Wiskunde & Informatica (CWI), The Netherlands

Prof. Frank van Harmelen, Professor, Knowledge Representation & Reasoning, Vrije Universiteit, msterdam (VU), The Netherlands

Fredrick Heintz, Associate Professor of Computer Science, Linköping University, President Swedish AI Society, Expert Member High Level Expert Group on Artificial Intelligence, European Commission, Sweden

Prof. Enrico Motta, Professor of Knowledge Technologies, The Open University, United Kingdom

Observatory for Responsible Research and Innovation in ICT (ORBIT), United Kingdom:Margherita Nulli, ORBIT Project OfficerBernd Stahl, Investigator, Director of the Centre for Computing and Social Responsibility at De Montfort University Martin De Heaver, Managing DirectorMarina Jirotka, Investigator, Professor of Human Centred ComputingCarolyn Ten Holter, Marketing Officer Paul Keene, Online Director

Prof. Ingrid Ott, Chair in Economic Policy, Karlsruhe Institute of Technology (KIT), Member of the German Expert Commission on Research and Innovation (EFI) 2014-2018, Germany Dr. Raymond Perrault, Senior Technical Advisor, Artificial Intelligence Center at SRI International, United states Giuditta De Prato, Team leader / Scientific Officer, Digital Economy Unit, Joint Research Centre, European Commission

Prof. Zhenan Sun, Institute of Automation (IAS), Chinese Academy of Sciences (CAS), China Prof. Tieniu Tan, Institute of Automation (IAS), Chinese Academy of Sciences (CAS), China

Prof. Chuan Tang, Associate Researcher, Chengdu Library and Information Center, Chinese Academy of Sciences (CAS), China

Dr. Zhiyun Zhao, Director New Generation Artificial Intelligence Development Research Center, Party Committee Secretary, Institute of Scientific and Technical Information of China (ISTIC), China

Artificial Intelligence experts

Elsevier is a registered trademark of Elsevier B.V. | RELX Group and the RE symbol are trademarks of RELX Intellectual Properties SA, used under license. © 2018 Elsevier B.V.