/sci/ - Science & Math

evidence

>>11286677
Replicable results using the scientific method.
https://en.wikipedia.org/wiki/Replication_crisis

Why basedboys love science so much and yet they dismiss actual studies that prove incels are in fact right?

>>11286691
elaborate

>>11286677
If you hear someone use words that could be seen as unclear and they cannot explain their statements in clear language, that's pseudo science speaking. Unclear word usage = pseudo science.

>>11286677
When the scientific method isn't being used properly. No it's not just whatever subject you don't like/look down upon you social science haters

>>11286810
>the scientific method
Not a real thing. Read Feyerabend.

>>11286816
based

>>11286695
https://incels.wiki/w/Scientific_Blackpill

>>11286816
Criticising the application of the scientific method isn't the same as it not being real retard

>>11286695
male variability hypothesis

>>11286677
Thank you, science girl Orei Soerjosoemarno(, PhD)

>>11286677
It's just about what authority you put your faith in.

>>11286677
Being falsifiable is a good start

>>11286682
But theretical physicists dont use the scientific method, yet you believe them?

>>11287309
>theretical physicists dont use the scientific method
an important step in the method is to form a theory in the first place

>you believe them
invalid statement, the act of titling something a "theory" implies that we don't have the evidence to completely believe it

>>11287309
Grant farming isn't exactly a science, anon.

>>11287027
It's real in the sense that there is a particular thing called the "scientific method" taught to 4th graders. It's not real in the sense that it's not how actual scientists operate.

>>11287403
Ahh ok, thanks for confirming my suspicions you have no actual experience real scientific research, what was your degree in?

>>11287460
BS in mathematics/chemistry
4th year candidate in theoretical chemistry

>>11287472
Prove it

>>11287389
>the act of titling something a "theory" implies that we don't have the evidence to completely believe it
>t. doesn't know what a "theory" is

>>11287506
Not gonna dox myself for a 4channel argument when you could just read Feyerabend, sorry.

>>11286816
>Feyerabend
Wiki:
>He argued that no interesting theory is ever consistent with all the relevant facts.
I hope you see the problem here.

>>11286677
>What makes something pseudoscience or not?

https://en.wikipedia.org/wiki/Pseudoscience

>>11287389
gravity is a theory

>>11286677
A failure to comply with scientific methodology, reliance on conjecture, anecdotes.

>>11286691
No they aren’t. Incels are a cult.

Go outside and talk to women like your dad wants you to

>>11287389
>an important step in the method is to form a theory in the first place

That doesn't work. Let me demonstrate:
Here is a sequence of three numbers. There is a rule to it. You need to figure out the rule. You can post your three numbers and I will tell you if they conform to the rule or not. You can test as many sequences as you want to, but you have only one chance of telling me what the rule is. Let's see if you can succeed or not.
3 6 9

>>11286699
>classic meme collection

>>11287666
interesting how nobody's willing to try

>>11287666
>>11288505
Yes, yes very clever I read Yud's ""work"" too.
The numbers are in ascending order.

>>11286677
>What makes something pseudoscience or not?

Easy one
If it confirms the Mainstream political narrative its real science and the big and respected Journals take you

If it challanges the Mainstream political narrative they shut you down and you mostly dont even get to publish your findings

I left the academia because of this, fuck my PhD im a high school teacher now and all those pozzed dickhead in the Unis can sucks a fat one

>>11288689
kek let's see your attempt at a study which was rejected by "them"

>>11288689
High school teaching is easier; you can make your own stuff up as it suits you, and the students are powerless to protest.

>>11288693
I worked in a Team that replicated svensmarks Experiment on the influence of radiation on cloud Formation, added with a few tweaks and additions from our side.
We basically confirmed his theory and made a computermodel of those datasets that actually had a much better r^2 when compared to the past temperature datasets than any co2 based Modell ever achieved.

Literally each and every "credible" journal told us to fuck off or didnt respond at all
Fuck that i have a PhD in astrophysics
I am teaching physiks and math at a private school now
Im done with this crap

>>11288516
And the same problem happens with science when you form a theory first - you will never make experiments that are predicted not to work.

>>11286677
A pseudoscience can't actually be used to make predictions. For example, two people can both be diagnosed with autism without sharing any symptoms, so knowing that someone has autism doesn't help you make any predictions about that person. Therefore psychiatry is a pseudoscience.

>>11287666
What's your point? Pick a theory, extrapolate it until you find a contradiction or some fact within the theory that doesn't correspond to reality, find a new theory that doesn't have the flaws of the old theory, repeat. This is how science works.

>>11287666
6 9 12

>>11287666
3 2 1

>>11287666
1 2 3

>>11286677
Refusal to comply with either rational reasoning and/or empyric evidence

>>11288799
3 - 6 - 9 - 11,446.7
Amirite?

>>11287666
3 6 9 9 4.5 1.125 0 0 0 0 0 0...

>>11289076
The 1.125 is not supposed to be there.
I just had a brain fart and forgot what I was doing.

>>11287567
>gravity is a theory
no, it's not. gravity is an observable phenomenon.
what we theorize about is how it works, and what causes it.

>>11287666
>you only have one chance of telling me what the rule is
fortunately, that's not how the world works.
basically what this guy said: >>11288792

>>11286677
>politically correct
Science
>politically incorrect
Pseudoscience

>>11289137
"Gravity" is a theory.
"Gravitation" is the observable phenomena.

>>11288711
>Literally each and every "credible" journal told us to fuck off or didnt respond at all
>Fuck that i have a PhD in astrophysics
Go back to bed Henrik Svensmark

when its not true

>>11286677
>tfw econchad and have to deal with stemfags saying econ is a pseudoscience because muh replication crisis

>>11288711
Did you put your article on a prepub archive like arxiv? No? Fake and gay never happened fuck off. If you did get ignored it's because your paper looked like it came out of a 5 years old's asshole, there was no science in it at all.

>>11287472
There is no way you have any degree in STEM and are unable to understand the scientific method.

>>11289214
Actually, you have them switched around because the name of the theory is "Theory of Gravitation" with Newton's "Laws of Gravitation", and gravity is the noun associated with the readily observable phenomenon of what we call gravitation.

>>11289212
You go girl

>>11286677
Organic chemistry and anything beside classical physics is pretty much a pseudo science. Also sciences like psychology as well.

>>11288799
>>11288802
>>11288806
What's the point if anon above revealed the answer.

>>11288792
>>11289137
The point of the rule is to make you answer only when you're absolutely sure the rule is right. The other point is
>the rule is x*3, (x+1)*3, (x+2)*3, hundreds of experiments confirm it.
>your measurements must be wrong. It's one of the best tested theories out there, and it says it isn't possible. Check your experiment, find a mistake
>It has been tested into tens of millions. If there will ever be a new theory, it will only involve extremely high numbers. Seriously just fuck off. Your results are wrong.
>decades wasted, new generation steps in
>the old ones were idiots. The correct formula is X, X+3, X+6
>.....

Can anyone send this thread link to her? you can send throwaway email to hers (aureliavizal@aol.com)

>>11289425
>The point of the rule is to make you answer only when you're absolutely sure the rule is right.
and that's how we know you're missing the point

it's not about being absolutely sure that we're right
it's about the process by which we approximate the solution

https://chem.tufts.edu/AnswersInScience/RelativityofWrong.htm

>>11290778
Why?

>>11290778
do it yourself faggot

>>11286677

>>11290821
>and that's how we know you're missing the point
You are missing the point, there is no magic man who could tell you your theory is right or wrong, it's only about the experiments you decide to make. Once you guess, it's pointless to continue.

>it's about the process by which we approximate the solution

No, this is completely wrong. People have been outrageously wrong many times. Only rarely it's as simple as gradual refinement.

>>11286677
Tl;dr
Pseudoscience makes non-falsifiable claims, scientific ideas are those that can be falsified

>>11290825
Because the thoughts of it make my peepee hard
>>11290863
i'm pussy

>>11286682
According to a 2016 poll of 1,500 scientists reported that 70% of them had failed to reproduce at least one other scientist's experiment (50% had failed to reproduce one of their own experiments).[8] In 2009, 2% of scientists admitted to falsifying studies at least once and 14% admitted to personally knowing someone who did.

>>11287309
Some theoretical physics is fine, but other parts are complete nonsense.

For example, Dark Matter was invented to explain galaxy formation, but they recently found galaxies supposedly without any Dark Matter, so how do those galaxies exist? The simpler explanation is that Dark Matter DOESN'T FUCKING EXIST. There's not a single direct piece of evidence for the existence of Dark Matter, nothing but missing puzzle pieces that we cobbled together into an explanation for unexplained behavior of gravity. Any competing theories get shot down by scientists who rabidly defend Dark Matter.

Another example is String Theory which is complete LUNACY. How many bright young physicists have been lost down that fucking rabbit hole? How much funding has been wasted on research relating to it? It's straight up crazy pills stuff, with zero relevance to reality and no impact on the real world (other than slowing down progress in other areas of physics).

>>11291311
Just because we don't know what dark matter is from a particle physics perspective doesn't mean there's zero evidence for it, pic and link related.
https://en.wikipedia.org/wiki/Dark_matter#Observational_evidence

>The simpler explanation is that Dark Matter DOESN'T FUCKING EXIST.
A hypothesis where dark matter doesn't exist must then explain why we apparently see the effects of dark matter, what is it about the galaxies that do appear to have dark matter that makes them different from those that don't? Explaining this is easy for the DM hypothesis. It's an unexpected find because of accepted models of galaxy formation but no-DM galaxies are not inconsistent with the dark matter hypothesis.

>>11291311
> Dark Matter was invented to explain galaxy formation

Most galaxies require dark matter for their existence and behavior to be possible.

>The simpler explanation is that Dark Matter DOESN'T FUCKING EXIST.

No it isn’t. Just that some galaxies don’t have dark matter.

> There's not a single direct piece of evidence for the existence of Dark Matter

Aside from the fact that most galaxies can’t exist without it being present.

> Any competing theories get shot down by scientists who rabidly defend Dark Matter.

The only competing possibility is a new theory of gravity, none of which exist.

>>11291311
Also the experiment that earned the dude a nobel prize for proof of black matter was recently shown to be invalid.

Clearly written down methodology which makes sure you can repeat the experiment based on the methodology used. Then results should be repeatable using the methodology. If different results for the experiment are found they should be noted and the previous experiment may prove to give unreliable results.

Still personally I'd say it is very important that all the conditions and how the experiment was conducted are written down exactly as they should. I think when it comes to this, modern academic journals are archaic considering source code, software used, datasets and so on are incredibly important in saving time for other researchers to replicate experiments yet there is no incentive to include these at all. Which leaves to many papers an unfortunate air of ambiguity.

>>11291325
There's actually no evidence of darkmatter that hasn't been disproved to this day.

>>11291282
> 1,500 scientists reported that 70% of them had failed to reproduce at least one other scientist's experiment

If you read papers and you try to implement part of someone else's experiment (or multiple parts) into your own, you will quickly find out that a lot of papers do not enable you to replicate their experiment. Vital steps are left out, important math in a model is missing to replicate it, some lines can be interpret in multiple ways. Which leaves you to wonder 'is there some code or w/e available to this experiment?'. The answer to this would be no. So you just disregard trying to implement what the paper was talking about and go and find something else. A lot of these papers I guess are not repeatable. I think any field that employs modelling, machine learning algorithms, statistical analysis, survey based experimentation suffers from this. As its really easy to be slightly fraudulent in this, but also to inflate results.

>>11291403
but if darkmatter doesn't exist then relativity is bullshit an daddy Einstein was a hack.
The tribe can't allow that.

>>11291408
Most ML stuff is actually abnormally reproducible because there's a strong culture of sharing the code and of solidly describing the model. Also the model is generally made from common parts rather than novel parts which further helps.
A reason some work is not reproducible that you haven't touched on at all and yet is very prevalent in all fields, including ML, is simply number fudging and lies. In ML specifically, it's very common to modify the evaluation metric to give a good score for crap results in a subtle way. For example I had to deal with a paper that claimed a PR-AUC of 0.6 on string prediction, when in fact it was redefining precision as "errors over predicted length" and recall as "errors over real length". Thus, it was in fact measuring the model's ability to predict length, not actual performance in the prediction (average PR for HELLO vs HAAAA is (1/5)/(1/5) = 1.0).

>>11291421
Absence of evidence is not evidence of absence. We have no tangible evidence for dark matter and previous evidence was wrong but that doesn't mean dark matter doesn't exist.

>>11291403
Factually false

>>11291434
[citation needed]
http://vixra.org/abs/2001.0077
By the way.

>>11291448
>vixra
>>>/x/

>>11291454
Congrats, you have extra chromosomes!

>>11291448
>[citation needed]
No u

>Vixra PDF made using Word
kek, anyway having skimmed through there doesn't appear to be anything in there to counter the dark matter hypothesis.

>>11291429
> because there's a strong culture of sharing the code and of solidly describing the model

Yeah I wish. Was dealing a year ago with a project where I had to do some predictive analysis on past financial data of defaults. Most papers had no data, nor code available. The one paper I found made very critical errors in not cleaning dataset appropriately so the results were kind of quatsch since I got much different results when I took their code and actually cleaned the dataset properly.

> it's very common to modify the evaluation metric to give a good score for crap results in a subtle way.

It is. But usually it ties into how the dataset is used and methodology as well. Often you can not expect clear code from these kind of studies either, or they omit their data for privacy purposes (they claim).

>>11291533
>or they omit their data for privacy purposes (they claim).
That doesn't matter since you can try it on your own data and see if it works. In any case you wouldn't give a shit about a model that works but only on very specific data, the whole point of ML in the first place is that the method is general even if the trained model is not.
>predictive analysis on past financial data of defaults.
You read a finance paper that used ML, which is slightly different, I agree. The same is true in e.g. biology papers from labs with no proper ML knowledge.
>The one paper I found made very critical errors in not cleaning dataset appropriately
And the same applies to the aforementioned biology papers.

If that's what you were referring to in the first place then you are right.

>>11290968
okay, so in the end when all is said and done how do you sort the wheat from the chaff?

>>11291429
>Also the model is generally made from common parts rather than novel parts which further helps.

most of which have been around since the 90's or earlier. lecun and hinton alone are probably responsible for most of the stuff people think is cutting edge today. hardware advancement really does drive the work. the state of the art in AI research is literally defined by how much ram nvidia decides to include in the next GTXXX-MUHFUGGEN-BIG-DICK-OVERDRIVE-EDITION meme card. and each time they do, tons of research is obsoleted.

>>11291731

and by that time, all the real low-hanging fruit is picked by companies like google who have the data and resources to conduct large-scale experiments.

there are a lot of clever papers out there, and a lot of clever students. but i think AI is a very tough area to make serious advancements. there are a lot of papers but most of them aren't worthwhile. how could they be when big tech has such a wide lead on academia?

>>11291754
cont.

and this is how you end up at a conference among a crowd of people with 7+ years of education, all gawking slack-jawed at a deepfake of tia tequila on trumps body or some other stupid shit. because those are the sloppy seconds that big tech has left them.

>>11286677
Idiots
Science is no less a religion than sky daddy obsession by similar cohort of sycophants.

A solid 70 - 80 of what is accepted as scientific fact is just flat out fucking incorrect. Why do you think that despite hundreds of years of investigation - thru the “scientific method” that society is in such a world of hurt.
People have never been sicker, Stupider, wealth inequality greater, more children committing suicide, more people in prison (US), more drug addicts legal and illegal, horrific dystopian panopticon fascism, and a massive maybe 80% of people that can’t think their way out of a paper bag. Wanton stupidity in the masses complements of the Scientific Methods.

>>11286677
90% of science is pseudo science

>>11291806
What a dumb fucking opinion. Did you make all those numbers up, or can you use some kind of non-scientific method to prove them correct?

>>11291786
cont.

i suppose that's a little harsh though. there is a lot of good research too. but it's not quite the gravy train it's made out to be.

>>11291806
Based but also cringe.

>>11286677
I think mostly psychiaters idea about it.

>>11286677
If the basis for rejecting/accepting a theory is if it hurts my feelings or not, then it's pseudoscience bullshit.

>>11291731
>tons of research is obsoleted.
Rather the opposite, it's just the same shit but scaled up. Beside VAEs, dropout, LSTMs/GRUs (same shit), ADAM (or any other adaptive learning algorithm like adadelta, in the end it's all the same shit - it makes the lr a tons less sensitive, saving hundreds of hours of work, but doesn't otherwise lead to better results and usually the opposite) and GANs (maybe unsupervised pretraining but nobody's even been using that in forever either thanks to computational improvements), there hasn't been anything new all this time (OK, maybe neural ODEs, but they're not yielding anything yet - maybe they'll become super cool like GANs being useless originally but being the norm now). Most advances are just "we ran the same model but bigger and the results were nicer!" garbage.
The other axis of learning, reinforcement learning, is mostly the same too. It's just hillclimbing or dynamic programming, but you use non-deterministic exploration and simultaneously estimate the cost of moves and the best solution instead of assuming the cost of move (as in A*/dijkstra, and other such methods).
Sutton is right in this blogpost:
http://incompleteideas.net/IncIdeas/BitterLesson.html
The other way to look at it is that because these methods are general and scale, they shall endure; whereas methods that rely too much on domain knowledge falter as the knowledge becomes invalidated or as more complex knowledge is uncovered, because general methods can exploit data to automatically detect this knowledge without human effort. You just replace 40 years of research into 40 hours of pseudo-magic mathematical optimization.

>>11291754
Big tech (like google) is really not leading in terms of research, it's really just "we ran an academic's project on big data and it worked please give award" bullshit. They almost never provide anything new. The latest was DRAW by deepmind and that's literally it.
Most papers in any field is not worth reading, period.

>>11291884
For example, almost every peer-reviewed papers in frontiers, nature method, pnas, etc. are downright fraudulent. Bin Ma and his lab's work is almost exclusively fraud (except for PEAKSDB) yet look at the publication record. Lots of the big names in proteomics alternate between publishing papers of the type "nothing works in proteomics reeee everyone's making shit up" and making shit up (it's because of the publish or perish culture so they have little choice). Every pathway analysis paper is a sham, almost every covariation or phylogeny paper is a sham. Everything that talks about gene signatures is a sham (the concept itself is a sham).

>>11291884
>Rather the opposite, it's just the same shit but scaled up.

this is sort of what i'm getting at.

>>11287516
>Doesn't know how to censor sensitive info
Nice larp bucko

>>11292352
What, you want a picture of my undergrad diploma? That's useless -- undergrad programs don't teach you science. As I stated, I don't have my PhD yet.

>>11292365
>What, you want a picture of my undergrad diploma?

Yes because I doubt you have even an undergrad degree.

The mental masturbation of philosophers around whether the scientific method is "real" or not does absolutely nothing to discredit its utility. It's a pity that you value such meaningless assertions.

Can anyone send this thread to her? You can DM her on twitter or send a throwaway email to aureliavizal@aol.com

>>11292828
nigga we dont need another schizo just stop

>>11286681
prove it

>>11286677
not following 'muh scientific method' or if it's a study that you don't like

>>11292842
I'm not schizo just need a little help right here

>>11286677
It depends on whether or not is is politically useful to call something pseudoscience

>>11286677
Real science tries to disprove its claims as much as possible.

>>11294338
after studying philosophy of science, I can confirm this is correct.

>>11286677
Whether Reddit believes it
A safe heuristic is to assume the opposite of whatever Reddit says
This will lead to some cookey beliefs like flat earth, but it will generally stand you in good stead.
For example, how many of us are actually going to space any time soon?
You may as well just believe the opposite of Reddit and live a good life as a fundamentalist Christian, racist, anti-gay, Trump supporting, large tech CEO, white male, Nazi, feminist, flat earther. It’s called creativity.

>>11286677
>What makes something pseudoscience or not?

There is no Pseudoscience just science
there are also no conspiracy theorys just theorys

science is literally all about some people saying stuff and other people proving it right or wrong, claiming there is some kind of authority that determines what is legitmate and what not kills the spirit of it

>inb4 being swarmed by basedboys
yea bring it on i am ready

>>11295513
She's not soyboy she's a qt 3.14

>>11286810
>scientific method
Please do explain what the "scientific method" is exactly. You could also not throw around meaningless buzzwords like a retarded pseud.

>>11290778
report this underage retard

>>11295781
You try as hard as you possibly can to prove whatever it is you think is correct wrong. Encourage others to. Maintain that you do not necessarily know that it’s correct if it hasn’t been proven wrong yet.

>>11291346
dark matter is a COPE for a flawed model.
>"this can only work if we invent a hidden invisible undetectable variable that acts in precisely the ways that our model works short"

>>11286677
Thoroughness, objectivity (involving agenda-pushing or subjective language is a big red flag), and reproducibility. All the things big daddy academia lacks.

>>11289338
>the scientific doesnt actually exist in practice, worst case it's a myth
>omg you dont understand the scientific method!
im not him but what the fuck? can you read, you mouthbreathing sack of slavish meat?

>>11286701
Define unclear.

>>11291408
>I think any field that employs modelling, machine learning algorithms, statistical analysis, survey based experimentation suffers from this.
so basically "climate science" then?

>>11293888
>>11292828
empty my testicles and I will subscribe to your IG

>>11286677
>What makes something pseudoscience or not?

reproducibility

>Science
Things I agree with
>Psuedoscience
Things I disagree with (also vaccines)

>>11286695
Female hypergamy mating strategy

>>11286677
Science - Jewish
Pseudoscience - Gentile

>>11299092
I don't have ig. I just want someone to send this thread to her. Can you help me?

>>11291346
>Aside from the fact that most galaxies can’t exist without it being present.
A - "Why do objects fall into the ground?"
Q - "Because there are invisible strings pulling them down!"
A - "What proof do we have that strings actually exist?"
Q - "The fact that objects fall into the ground!"

>>11286885
All of this is well known common sense. Women are attracted to strong bad boys.

>>11286885
>Making Tinder profile and my ex wants to help
>write up this nice bio about my interests and personality with pictures of me having fun
>show her
>”Anon, delete that and just show a picture of your muscles”
>TFW it worked

>>11286677
I heard that Nietzsche used to say that once you close a department and fire everybody, only few of your ex-employees will keep on doing science by themselves - those few are the true ones, all the rest are pseuds.

>>11301495
he is not wrong.

>>11286682
Psychology isn’t a science

>>11290875
Holy fuck my sides
Thanks anon

>>11290968
>U cant kno nuffin

>>11301963
> thinks replication crisis doesn't take place in his own field
> didn't have another thought about why it may take place
https://www.youtube.com/watch?v=B5dvUVuwcxg

>>11302902
The point is you need to be careful about what you think you know, and try to test not only if what the theory predicts actually happens, but also if what the theory doesn't predict doesn't happen.

>>11301507
>BUT NIETZSCHE INSPIRED NAZIS NOOOO YOU CANT AGREE WITH HIM YOU CIS HET PIG NOOOOOOO