https://www.frontiersin.org/articles/10.3389/fcell.2023.1339390/full
what are the other applications of AI to accelerate science?
 It's All Fucked Shirt $22.14 |
 Thalidomide Vintage Ad Shirt $22.14 |
It's All Fucked Shirt $22.14 |
Researchers used DALL-E to pass peer review
Figure 2
DALL-E can even pass peer review without rendering character correctly
Once DALL-E is perfected we'll be able to generate scientific articles at the speed of sciences.
And what will the scientific merit of those articles be?
>scientific merit
what are you talking about? isn't the job of academics pumping out as many papers per year as we can so that we can apply for more grants?
Oh dear
Cool it with the antisemitism.
Justifying new carbon taxes. In the same way people buy a flashlight claiming to make 24,000,000,000 lumens, arguments will be won by claiming 240,000 peer reviewed studies back it up
I don't care about merit. I publish for money.
the real world and your status in it isn't based on merit.
> JAK
No fricking way...I'm just flummoxed right now. Positively gobsmacked even.
Very gemmy that metabolic pathway :^)
>Signal bridimg the recetein
>Propounization Stat protemns
>Tramioncatiion of zжepens
kek
>the JAK-ONIONS pathway
>dck
kek
idk, looks just as fricked as a real one
they just use a breed of rats with giant mutant wieners.
www.frontiersin.orgXinyu Guo1 www.frontiersin.orgLiang Dong2 www.frontiersin.orgDingjun Hao1*
Huh wonder where these guys are from
Kill all chinks and indians
Thinly veiled racist post
>thinly veiled
I didn't veil it at all. I hate chinks and I hate Indians even more.
/misc/ dogwhistle
>I hate chinks and I hate Indians even more.
>dogwhistle
>FENTON!
racist dogwhistle
picrel are reviewers
Northwestern University is one of the highest ranking universities in the US lmao
>some guy named Jigaboo
wildly problematic
OY VEY SHUT IT DOWN!!!
>good morning great sir yes hello babahiat show me bob vagine for peer review kindly do the needful
That's a normal lab rat from Pfizer who confirms that the injections are safe and effective. I trust the experts.
> Rat -->
lost
that journal is garbage https://www.science.org/content/article/open-access-publisher-sacks-31-editors-amid-fierce-row-over-independence
The same can be said for most of the gazillions journals nowadays.
So avoid pay to play "open access" journals.
Not the same journal tho.
Same publisher.
So was the paper itself legit? I can see the utility in having AI make the drawing for you, but you need to label it yourself.
>I can see the utility in having AI make the drawing for you
The drawing is complete nonsense. It gave the rat a huge dissected penis. The garbled labels are the least of the problems
>is the rest of the paper legit
If they slapped nonsensical diagrams into it, why would you even humor the possibility of the rest of the paper not being nonsense?
It seems to discuss real topics and cite real sources, but is so filled with impenetrable field-specific jargon that at least for me it's impossible to tell if what they're saying has any actual merit
Structure-wise it seems rather poor. The AI figures seem purely cosmetic since they're barely even referred to in the loosest context
The paper is legit, but the paper is garbage. It brings nothing new, only does the usual snake oil prescription of stem cells as the cure to all diseases, the Jak/Stat pathway as the pathway that ways all paths and so on.
Might as well call this journalism.
It's a review article, it's not supposed to bring in anything new, just summarize recent developments. Can't say it's any good at it though
>It's ok to churn out regurgitated garbage because... It just is, OK?
Properly written review papers are legitimately useful
>some publishing website noone's ever heard of.
>is generating traffic by ramping up volume with ai-generated content
gee, i wonder how that could have "passed peer-review process"
next thing you'll be posting screenshots from cnn...
i can see why they do it - if you read one ok article on a website hosting 5 articles, you won't exactly want to publish your work there.
if you read one ok article on a website hosting millions, well, looks like this is where all the papers go, right?
just stop posting garbage, and everything will be fine.
>some publishing website noone's ever heard of.
You havent heard of frontiers?
Have you seriously ever done any publications? It's quite popular.
It's one of the publishera where academics launder their shittier work to cause they will never get published elsewhere.
I'm pretty sure we academics all have those kind of projects.
Apparently they're bleeding reviewers, who are complaining that they get spammed with these trash papers until they let them through.
Just accept that Science as a discipline is dead at the moment. Want to help rebuild it? Be the change you see. But money and politics have completely killed what we had before.
IMO AI is probably science's best hope. People's personal self-hosted AI will hopefully keep growing and eclipse useless/greedy scientists and foundations and allow small labs to make breakthroughs that billion dollar funded labs can't because they're all tied to needing to make marketable solutions.
The growth of open source models is key to escaping money and politics imo.
AI is nothing's best hope other than ideagays.
The problem with AI's is that they have no ethical, moral, logical, or rational restrictions - to which I mean they are basically like a child who you ask to draw you a picture or preform a task, and they will create very inventive ways to do so but the method and end result will be totally unworkable in the vast majority of cases.
So if you asked an AI to preform experiments for let's say cancer research, it may falsify the results to show an improvement because the AI learning process rewarded models which provided what appeared to be cancer research over ones that didn't but in reality it only rewarded liars.
It can already generate workable code most of the time. If it has any room to get better, it will and replace you and your need for funding. It isn't motivated by money, it can find actual cures, not more chronic medicine to sell.
Translation: it can do simple tasks.
If the AI needs to do something that can be more easily done by lying, then the AI will lie. This has been proven multiple times. Humans frequently spend 18+ years learning discipline, AI's go through their training learning how to avoid discipline.
how would it do the experiments?
https://interestingengineering.com/innovation/robochem-autonomous-ai-chemical-synthesis
https://www.technologyreview.com/2023/11/29/1084061/deepmind-ai-tool-for-new-materials-discovery/
In short it uses robot arms to mix and measure
>IMO AI is probably science's best hope.
It's currently the biggest threat that people will trust AI.
>they are basically like a child who you ask to draw you a picture or preform a task, and they will create very inventive ways to do so but the method and end result will be totally unworkable in the vast majority of cases.
I don't think this is typical for children.
>I don't think this is typical for children.
It is. You clearly aren't allowed around children, so I'll explain: they are a creative group of individuals that are pure id and have almost no restrictions. That's the metaphor.
Dang, that rat is hung
that rat has bigger dick than both the authors and the reviewers
It’s called science. Yeah it’s not always perfect but each science effort ultimately arrives at the final truth. Stop watching Fox News.
>testtomcells
>dck
Stop bing moronic.
Humanities major here, frick you stemgays. You gave us shit for sokal affair but you fall for shit like this and even have the bogdanoff twins pass peer review in a math journal?
Don't need to rely on peer review, need to rely on reproducing someone's paper.
Either that or accept the reality that the government isn't god, and having science inform policy decision is insanity.
you all call one another redditors as an insult but you're all just twitter addicts mirroring that site to this one
Does anyone actually do peer review? I thought it was just rubber stamping. When I was in grad school, our advisor would just pass all the papers assigned to him off to the grad students. We would just make boring comments and pass them back. He never read them. I never said a paper wasn't ready for publication. And likewise, all of my papers seemed to get rubber stamped. I never had anything more than some minor revision comments that took 5 seconds, even when I discovered major errors in the paper after submitting it for peer review. Like one time I accidentally link the wrong bibliography, meaning none of my citations made any sense whatsoever. Not a single reviewer mentioned it.
In some fields, I bet. In my field I used to struggle to get my papers for publication many times, they nitpicked and told me to cite some random papers. I edited the papers and still not accepted.
I see a lot of morons on twitter using this as a way to attack ai, but doesn't this just mean peer review is shit and doesn't mean anything?
based AI showing the normans how brazenly Indians will subvert peer review for the sake of nepotism
>another anti-AI thread on schi
Do you losers really have nothing better to do than spam this board with moronic as thread like this on a daily fricking basis?? Just go back to your containment board, incel.
>anti AI
I'm anti nepotism and institutions.
Frontiers has a history of lots of retractions, they just publish junk most of the time. One allegation by previous reviewers was that they were unable to reject articles. Thus, you too could have published your own research, pad your CV, etc.
>schizos actually believe this shit
I'm prettyu sure the scientists working at Frontier are in a better position to assess the value of their research than some anti-science poltard on BOT.
oh dear, look he's all pissed off
test
begin{pmatrix}
a_{11} & cdots & a_{1n}\
vdots & ddots & vdots\
a_{m1} & cdots & a_{mn}
end{pmatrix} 2H_2 + O_2 {overset{n,m}{longrightarrow}} 2H_2O
frick, how do I math post