Shulou(Shulou.com)11/24 Report--

This article comes from the official account of Wechat: back to Park (ID:fanpu2019), author: lobule

Recently, a study published by Nature, through a statistical analysis of tens of millions of papers and tens of thousands of patents, shows that the number of research papers has soared in recent years, but the growth rate of subversive discoveries and technological innovation has declined sharply. But we do not know the exact reason and how to improve it.

Write article | lobule

There is no denying that major scientific discoveries and technological breakthroughs will completely change our lives, subvert our perception of the world and change the operation of society. With the continuous growth of the scientific research team, the number of scientific research papers published and technology patents are growing rapidly, adding "new fuel" to promote the productivity of scientific research and technology. However, many scientists are not optimistic. On the contrary, they remain calm and objective, always pay attention to the direction of scientific research development, and also throw cold water on the public from time to time, telling everyone that the current development of science and technology may not be optimistic.

Conservative or subversive? In 1996, John Horgan, a senior contributor to Scientific American magazine, interviewed many famous scholars and showed scientists' deep thinking about the future of science in his book the end of Science (The end of science). The book asks many interesting questions about different disciplines, one of which is a direct attack on the major crisis hidden in science as a whole since the 21st century: has today's science declined to the point where it can only answer trivial questions and repair existing theories? [1]

A paper published in the journal Nature at the beginning of 2023 gave a near-affirmative answer to this question. To understand the degree of development of science and technology, papers and patents are the most intuitive and concise criteria. Russell J. Funk from the University of Minnesota (University of Minnesota) and two colleagues defined two types of breakthrough scientific discoveries and patented technologies: the first is to improve the existing scientific knowledge base and play a consolidated role (consolidation); the second is to subvert the existing scientific understanding, thus promoting the discipline to develop in a new direction and play a subversive role (disruptive).

So how to measure the role of consolidation or subversion? The author judges that papers and patents with a consolidated nature are more likely to cite previous research and may inherit existing knowledge, while for subversive research, subsequent studies are less likely to cite previous articles. On this basis, the team put forward a new index: CD index to represent the consolidation or subversion tendency of scientific and technological achievements. The index ranges from-1 (consolidation) to 1 (subversion) and is limited to five years after the year in which each paper / patent was published, expressed as CD5.

The team analyzed 25 million papers recorded on Web of Science (WoS) from 1945 to 2010, 3.9 million patents recorded by the United States Patent and Trademark Office (USPTO) patent browsing database from 1976 to 2010, and four other additional databases: JSTOR, the American physical Society Proceedings Library, Microsoft academic Scientific Literature and patents, and PuBMed, a total of 20 million papers. Through the CD5 index, clear sinking curves reveal the direction of science and technology that mankind is proud of since the middle of the last century: on the whole, new scientific knowledge and technology are becoming less and less subversive. Among all the disciplines examined, the CD5 index of papers / patents fell by 91.9% between 1945 and 2010. Since the 1980s, the decline in the CD5 index in the life sciences, biomedicine and physics has been moderate, while the decline in social sciences and technology has been more significant and lasting. During this period, for example, the CD5 index in medicine and medical technology declined by 91.5 per cent.

Fig. 1 the changing trend of CD5 index of achievements (papers and patents) in the investigated disciplines from 1945 to 2010. Picture source: reference [2] words and sentences in papers / patent documents are also used to verify the team's point of view. Subversive papers and patents tend to invent new words and new terms, so as to establish a new scientific paradigm to distinguish it from the existing one. Moreover, a subversive reduction may also lead to a reduction in the diversity of words used in science and technology (the study analyzed the words used in the title of the paper).

In addition, the choice of verbs in the literature has also changed. with the passage of time, verbs related to "create", "discover" and "perceive new things" are used less and less frequently. On the other hand, verbs related to "improvement", "application" and "evaluation of existing things" (with consolidation meaning) are used more and more frequently. From the perspective of literature language model, the team also proved that with the passage of time, our scientific discoveries and patented technologies have become more and more conservative, with the main purpose of consolidating existing knowledge rather than subverting and creating new fields.

Figure 2. The changes of the top 10 verbs most frequently used in the papers and patent titles published in the 1980s and 2010. The verbs commonly used in the 1980s to create, discover and perceive new things were almost completely replaced by verbs with meanings such as "improvement", "application" and "evaluation" by 2010. Photo Source: reference [2] this paper is on the cover of the first issue of Nature in 2023. In related news reports, other scientists also expressed their views on this. Dashun Wang, a computational social scientist at Northwestern University in Evanston, Illinois, said: "it's nice to see this [phenomenon] documented in such a meticulous way. They look at the problem in 100 different ways, and I find it very persuasive in general." [3]

Figure 3 Net loss, net loss. Photo Source: Nature large input, low return "Natural" thesis breaks our previous optimistic filter about the subversive nature of scientific research with solid data. in fact, as early as many years ago, the academic community has noticed another problem: we have trained more and more scientists, the state and institutions have invested more and more money to sponsor scientific and technological achievements, and more and more papers have been published. However, the productivity of scientific research has not "risen all boats".

Figure 4. The report "current situation of Scientific Research Productivity" gives examples of the trend of scientific research productivity in the United States in the past 80 years. Photo Source: "status of Scientific Research Productivity" in 2018, Merck commissioned the Oxford Institute of Economics to conduct interviews and surveys in organizations of different sizes and types in five industries in seven countries around the world, and reviewed the existing literature. in 2021, the report "status of Scientific Research Productivity" was published [4]. The Oxford team defined scientific research productivity as the relationship between input and output of scientific research activities. In addition to basic research and applied research, the team added experimental development to reflect the process of transforming scientific research results into commercial products, which are closely related and complement each other. In addition, the team uses the number of papers published and the number of patents granted as the measurement criteria, which illustrates the changing trend of scientific research productivity to a certain extent.

First of all, the number of papers published around the world has increased significantly over the past half a century. In 2018, the Natural Science Foundation of America reported that from 2008 to 2018, the number of peer-reviewed scientific and engineering journal papers and conference papers increased at an average annual rate of about 3.8%. In the same year, the International Association of Scientific, Technical and Medical Publishers (STM) calculated that the number of peer-reviewed journals has grown steadily at an annual rate of 3.5% since the 17th century, reaching a growth rate of 5% per cent per year from 2008 to 2018, with an average of 1.5 million to 3 million articles published each year [6].

Closely related to the increase in the number of papers is the increase in the number of researchers. According to the latest statistics of UNESCO [7], from 2014 to 2018, the number of global full-time work equivalent (FTE) researchers increased at a rate of 13.7% (13.7%) than the global population growth rate (4.6%). As of 2018, there were a total of 8.854 million FTE researchers.

Figure 5. In the United States, for example, from 1997 to 2019, the number of researchers and the number of papers published in the United States developed synchronously, showing a positive trend as a whole. Source: the abundant number of researchers and the publication of papers in the current situation of Scientific Research Productivity is an important pillar to promote scientific research productivity, while the other pillar is the investment of sponsorship funds. Taking the United States, which has the largest funding for scientific research in the world, for example, the American Association for the Advancement of Science (AAAS) has counted the trend of total R & D spending in the United States from the 1950s to 2019, which is growing rapidly. (see figure below) [8] in recent years, China has also caught up: from 2016 to 2021, China's R & D spending grew at an average annual rate of 12.3%, which is significantly higher than that of the United States (7.8%). The total national investment reached 2.8 trillion yuan, and maintained double-digit growth for six consecutive years. [9]

Figure 6. According to the statistics of the official website of the American Association for the Advancement of Science (AAAS), the trend of total investment in national R & D funding in the United States in the past 60 years, the source: the high input of human and financial resources in the scientific research community has led to a continuous increase in the quantity and productivity of scientific research papers, but many signs show that the increase in quantity has not led to an improvement in quality: a variety of journals have reduced the difficulty of publishing papers. "Research inflation" has led to the emergence of more low-quality research, other researchers are often unable to reproduce the experimental results of low-quality research, and the proportion of withdrawn papers due to errors and academic misconduct is also increasing.

The publication of a large number of papers not only can not reflect the real situation of scientific development, but may also hinder the development of science. A study published in the Proceedings of the National Academy of Sciences (PNAS) in August 2021 analyzed 1.8 billion citation records of more than 90 million papers between 1960 and 2014, revealing significant differences in citations: many papers were published, but a large number of cited papers were only a few. The more papers published annually in a field, the more papers cited in the previous year will still be highly cited in the next year, thus attracting the attention of most researchers and ignoring the newly published papers. This is the phenomenon of "class solidification" in the scientific research community. And the team also reached a conclusion similar to the cover paper of Nature, and worried that conservatism could easily lead to the phenomenon of "monopoly" in scientific research. [11]

On the other hand, at the two levels of application and experimental development, which are more closely related to actual production and economic benefits, the Oxford team uses the number of patent applications and grants per million US dollars per euro of R & D expenditure as the measure of scientific research productivity (Note: the report acknowledges that the method does not take into account patent quality, influence and practicality, so the conclusion is more illustrative than deterministic). Found that except China and Japan The number of patent applications and authorizations in many countries is basically the same or declining, indicating that the corresponding application R & D productivity is stagnant or even declining.

In the electronics industry, for example, although Moore's Law predicts that the number of transistors on computer chips doubles about every two years, this is not the case. According to an analysis published in the American Economic Review (American Economic Review) in 2020, compared with the 1970s, it is 18 times more difficult to achieve exponential growth in chip density than it was more than 50 years ago, while R & D productivity is declining at a rate of 6.8 per cent a year. At the same time, the new generation of electronic technology requires more R & D resources, iterative update costs increase by 35%, and the next-generation process costs of "fab-free" semiconductor companies will increase by 60% [13].

Figure 7. From 1930 to 2000, the number of researchers in the electronic chip industry continued to grow, but the corresponding scientific research productivity was declining at a rate of 6.8% a year. Source: 10.1257 / aer.20180338 reasons for the decline of scientific research productivity in the face of the complex current situation of the subversive decline of scientific research and the decline of scientific research productivity, academia is still exploring the underlying reasons, and different experts and scholars have put forward a variety of possible reasons. For actual scientific researchers, many questions may be platitudes.

The theory of "scientific research burden" (the burden of research) [14] is one of them, that is to say, with the development of science today, in order to fully and thoroughly master professional fields, from student days to becoming researchers, we must constantly learn a large amount of relevant professional knowledge and follow up new knowledge in time. Scientific knowledge as a whole shows an explosive growth momentum, and the time line that needs to break through the boundaries of professional fields has also been lengthened. Not to mention really subverting a discipline.

Scientists and inventors who are overwhelmed by the wave of expertise also tend to be "confined to vertical areas of expertise and get caught up in narrower research topics, resulting in being busy consolidating improvement rather than thinking about subversion" [13]. This is the opinion given by Michel Park, the lead author of this year's Nature cover paper. The author also acknowledges that the growth of scientific knowledge has a positive impact on the subversion of the paper, but a negative impact on patented inventions. Therefore, in order to promote the emergence of disruptive scientific and technological achievements, the team recommends that scientists read more widely and expand their understanding of cutting-edge knowledge.

In view of the fact that the current scientific research is based on the accumulation of hundreds of years of achievements, constantly pushing the frontier of the discipline further, the theory of "low-hanging fruit" (low-hanging fruit) comes from economics, that is, there are always some fruits on a fruit tree whose position is low, so they can be picked quickly, and the higher the fruit position, the more difficult it is to pick and the greater the cost. This model is also applicable to explain the development status of the scientific research community [15]. Standing on the cornerstone of existing scientific knowledge, a scientific point of view with a little innovation that can consolidate the existing cornerstone is like a "drooping fruit". It is easier to win, but the competition is often fierce and will be picked less and less. The rest are the fruits hanging high, that is, scientific viewpoints that can subvert the existing achievements of a discipline, or guide the creation of a whole new field within the discipline, which will only become more and more difficult to extract, which not only requires the strength of the researchers themselves, but also involves many other practical interests, such as team size, funding applications, and so on.

The above-mentioned factors basically start from the scientific research content itself, while more problems appear in the scientific research environment itself.

The report "current situation of Scientific Research Productivity" directly summarizes four main negative factors of scientific research productivity: the administrative burden caused by the increase in the size and professionalism of scientific research teams; the "thesis-only" scientific research evaluation system; the imbalance between cooperation and "outsourcing"; and the slowdown of financial support in many countries, and points out that the financial support of the United States government in experimental development is lower than that of basic research or applied research.

How to go in the future of science? Today, scientific development seems to be moving towards the goal of consolidating scientific knowledge. Compared with the last century, the pace of progress and productivity have also slowed down. It can be said that we have entered a relatively stable era of incremental science. "A healthy scientific ecosystem includes both consolidation and improvement of previous work as well as new subversive discoveries, but the nature of the research is changing," Funk told Nature. "as incremental innovation becomes more common, it may take longer to make key breakthroughs that clearly promote scientific development in the future."

However, some scientists believe that we should not make an extreme choice between incremental science and subversive science. Wang Dashun believes that subversion is not essentially good, and incremental science is not necessarily bad. The healthy combination of the two can sometimes bring great surprises. For example, the first direct observation of gravitational waves is both revolutionary and the product of incremental science. John Walsh, an expert on science and technology policy at the Georgia Institute of Technology in Atlanta, says the ideal is to combine incremental research with subversive research healthily. "in a world where we focus on the validity of our findings, more repetition and repetition can be a good thing," he said. " [3]

Nowadays, with the vigorous development of artificial intelligence technology, artificial intelligence authors have been added to some papers, which can easily learn vast amounts of existing knowledge, which may bring new changes to our incremental science era.

references

[1] John Hogan (John Horgan) The end of Science (revised Edition) [M]. Beijing: Tsinghua University Press, 2017

[2] https://www.nature.com/articles/s41586-022-05543-x

[3] https://www.nature.com/articles/d41586-022-04577-5

[4] https://www.merckgroup.com.cn/content/dam/web/corporate/non-images/scientific-research-productivity/Scientific-Research-Productivity-White-Paper-CN.pdf

[5] https://ncses.nsf.gov/pubs/nsb20206/

[6] https://www.stm-assoc.org/2018_10_04_STM_Report_2018.pdf

[7] https://www.unesco.org/reports/science/2021/en

[8] https://www.aaas.org/programs/r-d-budget-and-policy/federal-rd-budget-dashboard?adobe_mc=MCMID%3D09531021069138174360329634836522590377%7CMCORGID%3D242B6472541199F70A4C98A6%2540AdobeOrg%7CTS%3D1673799539

[9] http://www.gov.cn/xinwen/2022-08/31/content_5707595.htm

[10] https://www.pnas.org/doi/10.1073/pnas.2021636118

[11] https://www.linkresearcher.com/information/7da728b6-9a5d-4f40-be4e-a779acb3fc09

[12] https://www.aeaweb.org/articles?id=10.1257/aer.20180338

[13] https://ide.mit.edu/wp-content/uploads/2017/12/IDE-Research-Brief_v1217.pdf?x96981#:~:text=Across%20a%20broad%20range%20of,will%20slow%20in%20advanced%20nations.

[14] https://www.zmescience.com/science/news-science/is-science-going-through-an-existential-crisis-there-are-more-research-papers-than-ever-but-innovation-is-sorely-missed/

[15] Cowen, T. The Great Stagnation: How America Ate All the Low-Hanging Fruit of Modern History, Got Sick, and Will (Eventually) Feel Better (Penguin, 2011).

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.