Home
Connexion
S'inscrire
Contenu d'apprentissage
Loading...
The Most Cited Science Papers of All Time - Video học tiếng Anh
Pratique d'écoute
Pratique d'écoute
/
Video
/
SciShow
/
The Most Cited Science Papers of All Time
The Most Cited Science Papers of All Time
Choisir le mode d'apprentissage:
Voir les sous-titres
Choisir le mot
Réécrire le mot
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Sous-titres (217)
0:00
This is the most popular science of all time!
0:04
Makes sense that SciShow is covering it.
0:06
Pop sci is kind of our thing.
0:08
But I mean something different when I say this video is about the most popular science ever.
0:13
I mean it’s the most popular science according to scientists.
0:17
…Based on data!
0:18
Here are the most cited publications of all time, from five big fields of science.
0:23
And I bet you’ve never even heard of them.
0:26
[♪ INTRO]
0:29
In April of 2025, a paper published in the journal “Nature” ranked academic
0:34
publications based on how many other papers referenced their work with a citation,
0:38
going back to the beginning of their databases.
0:40
So we went through the list and compiled the number one
0:43
most cited papers from five different fields.
0:46
And we’re going to count down to the most cited one, starting with materials science.
0:51
The fifth most cited paper on our list has 52,000 citations.
0:56
It’s all about a legend in the world of materials science: graphene.
1:00
Yes, one of the most popular publications in the
1:03
world describes a material you probably know very little about.
1:06
You’re more likely to be familiar with its cousin, graphite.
1:10
The stuff in pencils.
1:11
But graphene is actually much cooler.
1:14
It’s a two dimensional version of graphite.
1:16
It’s literally one atom thick, with its carbon atoms arranged
1:20
in a flat array of hexagons, like a drawing of a honeycomb pattern.
1:24
And that shape is so bonkers that back in the 1940s, many scientists didn’t believe it
1:30
could physically exist, since a material that thin would be extremely unstable.
1:34
But in 2004, researchers managed to make this impossibly thin carbon sheet IRL.
1:40
That’s when graphene really made a name for itself.
1:43
And it turned out that honeycomb shape was a game changer, making it way more functional.
1:48
Materials get people’s attention when they have specific properties that make them useful,
1:53
like their ability to react with light or electricity, their lightness, or their strength.
1:58
And graphene happens to check a lot of those boxes.
2:01
Graphene is an incredibly useful material because of its conductivity, thinness, and stability.
2:07
Let’s start with conductivity.
2:09
That quality gets people, and electrons, going because pretty much anything
2:14
involving electricity needs a conductor, or a material that electrons can flow through.
2:19
Usually conductors are metal, since the structure
2:21
of metals already has plenty of space for its electrons to all move together.
2:25
It’s like a crowd of marathon runners running along the same race course.
2:29
But metal has some limitations.
2:31
And graphene can do the same thing while being made entirely of carbon.
2:35
The secret is its thin honeycomb atoms.
2:38
That orientation lets electrons skate across the surface without interference from other atoms.
2:43
So it’s a great conductor.
2:45
But it’s also thin enough to be used in applications like nearly
2:48
transparent coatings and smartphone screens.
2:51
And it’s strong enough to reinforce the parts
2:53
of a tennis racket that take the most force when hitting the ball.
2:56
Because of all the uses graphene has today, the paper has garnered
3:00
a lot of attention from scientists over the years.
3:02
And that’s what makes this graphene review the most popular publication in materials science.
3:07
Now, you might expect that as this list goes on, the papers will get older.
3:11
They’ll have had more time to accumulate citations, after all.
3:14
But sometimes, a new topic bursts onto the scene with so much enthusiasm behind it
3:19
that it quickly moves up the ranks and surpasses everything that came before.
3:23
That’s what happened for the next entry in our list:
3:26
the 2015 paper “Deep Residual Learning for Image Recognition”.
3:30
This paper has more than 116,000 citations.
3:35
And while it’s not the first publication to
3:37
describe how deep learning could be used in the context of images,
3:40
it became a landmark paper in this fast-growing field by proposing a much easier way forward.
3:46
It suggested an improvement to the way we build image models.
3:49
And that’s wildly popular because we see the world in images.
3:53
So, naturally, they make up a lot of the world’s data.
3:56
Think about all of the images we take of space to explore it,
3:59
and the images we take of organs in the human body to diagnose people.
4:03
Overall, our brains are pretty good at understanding what’s going on in those images.
4:07
…But not so good at doing that for thousands of images, at least not fast.
4:13
And that’s where computing for images comes in.
4:15
Image recognition AIs are sort of modeled after the way our brains work,
4:19
but can interpret images significantly faster.
4:22
They’re layered models known as neural networks,
4:25
which “learn” patterns from lots and lots of examples.
4:28
Starting with an image, each layer does a little math and passes information down the
4:33
line to the next layer, and the next, until it generates the output we want.
4:37
For example, a label for objects in the image.
4:40
And models with lots of layers tend to perform better on tasks like image
4:44
recognition because of how complex images are.
4:47
There’s so much to learn that having all those
4:49
layers gives them more opportunity to identify various details.
4:53
But sometimes, when networks have tons of layers,
4:56
they can lose track of some information on the way.
4:58
Each layer of the network is transforming information.
5:01
With, say, 100 layers, it gets to be like a game of telephone.
5:05
To address that problem, the most popular paper in computer science
5:09
proposes grouping the model's layers into chunks.
5:12
Now the model can take shortcut paths from one chunk to the next.
5:16
And a model that uses these shortcuts also still
5:19
maintains the direct path of information through each layer.
5:23
We’ve just added another mechanism to make sure nothing gets lost.
5:26
They call this method residual learning.
5:28
And while image recognition isn’t its only use,
5:31
it’s especially helpful for complex images that require really deep neural networks.
5:36
For applications like self-driving cars and
5:39
identifying tumors on scans, we don’t want our neural networks losing any information.
5:44
And those innovative uses for deep residual learning just keep coming,
5:49
which is why this paper is the most popular in all of computer science.
5:52
Image recognition is all around us.
5:55
But not all of the top cited papers have their utility right in the title.
5:59
One example is the most-cited paper in physics,
6:02
titled “Generalized Gradient Approximation Made Simple”.
6:06
Spoiler alert, it was not made simple.
6:09
But enough physicists seem to have understood its value,
6:12
because it's racked up more than 174,000 citations.
6:16
That’s because this paper made it easier to solve the huge problem
6:20
of describing our tiniest components; the things that make up everything.
6:25
We’re talking about atoms.
6:26
They’re hard to describe because they’re just so dang small!
6:30
Let’s put it this way.
6:31
When you sit at the top of a slide, we know that gravity will pull you down,
6:36
even if you don’t get a push.
6:37
You have high potential energy at the top of the slide,
6:40
and will move toward a place where you have lower energy.
6:43
But atoms probably don’t go down slides like we do.
6:46
Although how cute would that be?!
6:48
Little atoms going down the slides at the atom playground?
6:52
At scales that tiny, matter starts acting …weird.
6:56
The mass of particles like electrons is almost nothing.
7:00
So if they were at the top of a teeny tiny slide,
7:03
we can’t say for sure that they’d get pulled down to the bottom.
7:06
And we can’t really ignore the way electrons in atoms move, because figuring that out is
7:11
critical to understanding how little things like drugs in our bodies work.
7:15
So physicists came up with the density functional theory of quantum mechanics.
7:19
This theory lets us estimate what electrons are doing in an atom, and how much energy they have.
7:24
And from that, we can figure out how they interact outside the atom as well.
7:28
Based on the charge of an atom’s nucleus and on other atoms nearby,
7:33
electrons might gather round an atom’s nucleus more densely or disperse.
7:37
And we can figure out when electrons disperse
7:39
or crowd together using the density functional theory.
7:42
But these calculations were not simple, which is probably where this paper’s title came from.
7:48
Each molecule is its own unique environment.
7:51
So to model electron density accurately, we needed to account for its specific arrangement of atoms.
7:57
That is, until this paper came along,
7:59
proposing a more streamlined way to approximate electron density.
8:03
Now, instead of figuring out the nuances of each environment,
8:06
we have an equation that depends only on physical constants.
8:09
That’s stuff like the speed of light, that won’t change for each situation.
8:13
So that kind of calculation just got a lot easier to do.
8:16
The outputs of these calculations have helped scientists understand
8:20
how current moves through semiconductors,
8:22
how to make better pharmaceuticals, even how to make more breathable fabric in your sneakers.
8:27
It might not seem as flashy as the pop science you’re used to, but there’s such a wide range
8:32
of applications that this publication is the most popular physics paper of all time.
8:37
We’re about halfway through our list of popular science.
8:40
So it’s time for a quick ad break.
8:43
Thanks to our Presidents of Science, binorthedrunkdwarf,
8:46
Charlie Stanley, and Harry Plumley for supporting this SciShow video!
8:50
In 2026, it’s not hard to get information.
8:53
The hard part is making sure it’s real information; not hallucinated by AI,
8:58
not disinformation that funds someone’s agenda,
9:00
and not well intended but ultimately incorrect misinformation.
9:04
At SciShow, we do the work of fact checking every single episode,
9:08
using peer-reviewed research as a primary source for everything we say, and maintaining
9:13
an AI policy that requires real people to be writing, editing, and presenting our videos.
9:18
And also require the visual in those episodes to not be AI generated either.
9:22
That kind of rigor costs more.
9:24
But it’s worth it because you know that the
9:26
information you get from us is held to high standards.
9:29
And we can only afford to keep up with those standards thanks to our incredible
9:33
Presidents of Science and other patrons, who support us at patreon.com/SciShow.
9:40
You might have noticed by now that the science that stands the test
9:43
of time is the stuff that’s incredibly useful.
9:46
Like an optimized method for studying proteins that’s used,
9:49
almost unchanged from 1970 to this day.
9:52
This paper has more than 250,000 citations with the catchy title,
9:58
“Cleavage of structural proteins during the assembly of the head of bacteriophage T4.”
10:03
But, uh, that title isn’t what made it so popular.
10:06
It racked up the citations because this paper introduces SDS-PAGE,
10:11
a method used to separate proteins by their mass.
10:14
Figuring out the size of molecules in our bodies is a major step in
10:18
characterizing those molecules so doctors and researchers know what they’re dealing with.
10:22
If you’ve ever done a genetic testing kit, for example, you’re looking for a variety of genes.
10:27
The DNA encoding different genes have different,
10:30
reliable sizes, which helps professionals identify them.
10:34
But proteins aren’t as straightforward as DNA.
10:37
Their charge and shape are a lot more variable.
10:40
So we need a little extra help to identify their size.
10:43
That’s what sodium dodecyl sulfate, or SDS does.
10:47
While this method was iterated on for a while, the most popular biology paper of all time is
10:53
the one that locked it in and demonstrated it on the protein envelope of a virus.
10:58
This version of SDS-PAGE is now a daily tool for labs diagnosing HIV, among other things.
11:04
It separates proteins to help clinicians identify those associated with the virus.
11:09
Which means that lifesaving techniques are the most popular biological science.
11:13
As they should be.
11:14
Finally, the most cited paper ever as of April 2025 is in the field often
11:20
called “the central science.” It’s chemistry.
11:23
This paper, published in 1951, has more than 350,000 citations.
11:30
That’s because it describes a method that has
11:32
become the basis for many tools that scientists use every day.
11:36
In labs across the world,
11:38
researchers are constantly measuring how much protein is in their samples.
11:42
Those could be samples of soil, blood, or even your protein shake.
11:45
I mean, you need to know if you’re going to meet your macro goals for the day.
11:49
Even though the method in this paper, called the Lowry assay,
11:52
is based on old technology, it’s very sensitive and provides consistent results.
11:56
So we still use it, often to fill in the gaps where newer assays won’t work.
12:01
And when I say “assay,” I just mean a lab test that tells you what’s in your sample.
12:05
This assay gets its utility from the Folin Phenol reagent, which is usually clear.
12:10
But under certain conditions, it can become oxidized, and turn bright blue.
12:14
The Lowry assay creates those oxidizing conditions by adding copper to a sample of protein.
12:20
Then, with more protein present, more of the reagent turns blue,
12:24
creating a more intense blue color in the solution when protein concentration is high.
12:28
The day when your science teacher brought out the
12:30
demonstration of color-changing solutions was always the best part of that class.
12:34
And it’s even better when the color tells you something about a person or environment’s health.
12:39
So even in the days of self-driving cars and virtual doctors’ visits,
12:44
this old and simple chemistry trick is still everyone’s favorite thing in science.
12:48
Maybe the most cited papers in history aren’t the most exciting discoveries to the average person.
12:54
But just like the foundation of a building,
12:56
the foundations of science aren’t always eye-catching and newsworthy.
13:00
Instead, they’re often solid and reliable,
13:03
stepping stones for thousands of other scientists to build with.
13:06
Whether they’re glamorous or not,
13:07
these papers have been integral to creating the world we live in today.
13:11
So now you can add us to the long list of people citing them.
13:16
[♪ OUTRO]
