Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The diffusion of scientific discovery

tchristos.com

Post reply on HN
Print view View original

11–16 of 16 posts

Re: The diffusion of scientific discovery

#11

Post by NoMoreNicksLeft »

Looking at their little toy simulation, one thing jumps out at me. Is there any particular reason that we should expect the labyrinth to not be pathological in its shape? Does the universe somehow owe it to us for it to be traversable?

One could, for instance, design the shape of the labyrinth such that most of the little jiggling dots not only can't make their way out, but also get trapped in a direction very far from the actual truth.

I hope that the random nature of the universe takes it a little easier on us than that.

Re: The diffusion of scientific discovery

#12

Post by bryanrasmussen »

jacquesm wrote: post #7
unclesaamm wrote: post #6

While I think simulations and models are interesting, this model is just way too simplified to give any real insight. Here's an example of a leading scholar in the field -- Allie Morgan -- developing a simple model on this exact question! https://epjdatascience.springeropen.com/articles/10.1140/epj... > Title: Prestige drives epistemic inequality in the diffusion of scientific ideas > Abstract: The spread of ideas in…

> We find that research from prestigious institutions spreads more quickly and completely than work of similar quality originating from less prestigious institutions. That's a bit of a leap though, isn't it? How to establish that the works are of 'similar quality'?

Well, this > As a result, well-known scientists tend to receive more credit than lesser-known scientists for work of comparable quality [27].

basically just references a bunch of other works.

One way of quality measurement is with the benefit of time I suppose - which would also be the way you would have the best metrics for idea diffusion.

If the future is already here just not evenly distributed yet it follows that what is evenly distributed now at some point in the past was unevenly distributed. With a good enough paper trail, which scientific publications provide, we could have a model of the distribution of these ideas that we believe, with the benefit of hindsight, are of similar quality.

Re: The diffusion of scientific discovery

#13

Post by teekert »

smitty1e wrote: post #2

I don't understand the need for a labyrinth. Why not simple particles in solution?

The walls are obstacles for moving on. Like storage and transport are obstacles for solar energy. But once we find solutions, we hit a next barrier. They can be large (ie not so much changes) or small (we get a series of follow up innovations, because energy has just becomes much cheaper and more abundant).

Re: The diffusion of scientific discovery

#14

Post by jacquesm »

bryanrasmussen wrote: post #12
jacquesm wrote: post #7

Earlier quoted context omitted.

> We find that research from prestigious institutions spreads more quickly and completely than work of similar quality originating from less prestigious institutions. That's a bit of a leap though, isn't it? How to establish that the works are of 'similar quality'?

Well, this > As a result, well-known scientists tend to receive more credit than lesser-known scientists for work of comparable quality [27]. basically just references a bunch of other works. One way of quality measurement is with the benefit of time I suppose - which would also be the way you would have the best metrics for idea diffusion. If the future is already here just not evenly distributed yet it follows that…

Sure, but if there is some kind of feedback loop in there it will mask out even the effects of time.

Some possible feedback loops:

Well known universities have more funds available and will be able to attract the best teachers away from the lesser universities.

Researchers may be willing to cite work from well known universities over similar work done at lesser universities.

The work done at other universities may be in a language that the researchers are not familiar with.

And so on. Quality in a vacuum is a very precise term but in the real world you'd have to discount for all of those factors.

In the tech world we have a similar example: Silicon Valley is 'where it happens' because it has a head start that is impossible to compete with which automatically attracts new talent and funding to a degree that it starves locations elsewhere resulting in an equality that dwarfs any qualitative effects that you may want to measure.

Re: The diffusion of scientific discovery

#15

Post by bryanrasmussen »

jacquesm wrote: post #14
bryanrasmussen wrote: post #12

Earlier quoted context omitted.

Well, this > As a result, well-known scientists tend to receive more credit than lesser-known scientists for work of comparable quality [27]. basically just references a bunch of other works. One way of quality measurement is with the benefit of time I suppose - which would also be the way you would have the best metrics for idea diffusion. If the future is already here just not evenly distributed yet it follows that…

Sure, but if there is some kind of feedback loop in there it will mask out even the effects of time. Some possible feedback loops: Well known universities have more funds available and will be able to attract the best teachers away from the lesser universities. Researchers may be willing to cite work from well known universities over similar work done at lesser universities. The work done at other universities may be…

ok sure but

>Well known universities have more funds available and will be able to attract the best teachers away from the lesser universities.

if we see that the work 20 years ago at a lesser known university was of the same quality as the work at a well known one, then I don't really get your argument about the money?

>Researchers may be willing to cite work from well known universities over similar work done at lesser universities.

which would really be confirming the argument that similar levels of quality get treated differently given their source.

Re: The diffusion of scientific discovery

#16

Post by unclesaamm »

jacquesm wrote: post #7
unclesaamm wrote: post #6

While I think simulations and models are interesting, this model is just way too simplified to give any real insight. Here's an example of a leading scholar in the field -- Allie Morgan -- developing a simple model on this exact question! https://epjdatascience.springeropen.com/articles/10.1140/epj... > Title: Prestige drives epistemic inequality in the diffusion of scientific ideas > Abstract: The spread of ideas in…

> We find that research from prestigious institutions spreads more quickly and completely than work of similar quality originating from less prestigious institutions. That's a bit of a leap though, isn't it? How to establish that the works are of 'similar quality'?

It's a simulation (following the theme of the OP submission), so these outputs are assumed to be the same quality. The point is even if you assume the quality is the same, the structure of the faculty hiring network alone will drive faster/further diffusion from prestigious institutions.
Post reply on HN