The Most-Cited Papers of the Twenty-First Century (nature.com) 13
Nature has published an analysis of the 21st century's most-cited scientific papers, revealing a surprising pattern: breakthrough discoveries like mRNA vaccines, CRISPR, and gravitational waves don't make the list. Instead, a 2016 Microsoft paper on "deep residual learning" networks claims the top spot, with citations ranging from 103,756 to 254,074 depending on the database.
The list overwhelmingly features methodology papers and software tools rather than groundbreaking discoveries. AI research dominates with four papers in the top ten, including Google's 2017 "Attention is all you need" paper that underpins modern language models.
The second-most-cited paper -- a 2001 guide for analyzing gene expression data -- was explicitly created to be cited after journal reviewers rejected references to a technical manual. As sociologist Misha Teplitskiy noted, "Scientists say they value methods, theory and empirical discoveries, but in practice the methods get cited more."
The list overwhelmingly features methodology papers and software tools rather than groundbreaking discoveries. AI research dominates with four papers in the top ten, including Google's 2017 "Attention is all you need" paper that underpins modern language models.
The second-most-cited paper -- a 2001 guide for analyzing gene expression data -- was explicitly created to be cited after journal reviewers rejected references to a technical manual. As sociologist Misha Teplitskiy noted, "Scientists say they value methods, theory and empirical discoveries, but in practice the methods get cited more."
Re: (Score:2)
Re: (Score:2)
Wrong century.
Re: (Score:2)
Or the summary.
Re: (Score:2)
I'm just guessing on that. I haven't read the fucking article.
If physicists were still trying to make direct progress on research from 1915 that would suggest the field was pretty stagnant.
I'm actually taking a mini-break from writing a web app that wraps around some ML at the moment. And I'm not referencing articles on how to write assembly, or even the specific structure of a TCP/IP packet.
... Of the twenty-first century? (Score:2)
C'mon its not even a quarter of the way though yet.
(The 21st century began on January 1st 2001 and runs til December 31, 2100 (inclusive)
The list is quite misleading (Score:4, Interesting)
The problem is that a lot of fields are dumped together, and that tells one nothing about the significance of the paper in that particular area.
For example, it is well-known that medical articles have very high number of citations, which isn't the case for papers in astronomy, and the reason is simply the size of the respective communities.
So, rather than rating on scientific significance, ranking articles like this one gets a rather meaningless convolution of the popularity of a topic, the popularity of a domain and the size of the peer audience.
A split by scientific area would have been more meaningful.
Re: The list is quite misleading (Score:2)
Re: (Score:2)
Some fields like CS think that a conference paper is worth as much as a traditional journal paper.
(ok this is a bit cheeky, mod accordingly:)
Not surprising (Score:4, Insightful)
I'm not surprised that methods papers are the most cited.
If somebody comes up with a useful technique, everybody who uses that technique is going to cite the paper.
meaningless, as noted in tfa (Score:3)
The most cited papers are ones that established some technique. It's like if every computer science paper cited Von Neumann for creating the stored-program idea.
CONVENIENCE SAMPLE (Score:1)
Unfortunately flawed data collection means none of this is likely true.
It's a failure to properly statistically sample that creates what's called a "convenience sample." Think of bringing a clipboard to a shopping mall just after school lets out and kids are coming by to get a coffee, enjoy an arcade, buy cheap stuff, etc. The data they provide is very easy to get (hence convenience) but is entirely different than sampled data across the entire day, bored housewives, etc.
This is the same problem here. Ci
This is well known (Score:2)
Papers that make wild discoveries usually close a field and so they don't necessarily get cited much.
Papers that create tools, datasets, methods will be cited by everyone that uses it.
At some point the most cited paper was the paper describing libsvm and it was gathering something like 4 citations per day.