machine learning: this place is not a place of honor

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > machine learning: this place is not a place of honor

«‹›24 »

animist: Aug 28, 2018

lmao

https://twitter.com/farbandish/status/1134485622352240640

# ? Jun 3, 2019 21:21

Adbot: ADBOT LOVES YOU

# ? May 11, 2024 21:33

power botton: Nov 2, 2011

people complain that facial recognition doesnt work on minorities but then when you try to do it all of a sudden everyone starts crying genocide.

# ? Jun 3, 2019 21:33

I HAVE GOUT: Nov 23, 2017

worthless people tend to complain a lot

# ? Jun 3, 2019 23:08

rotor: Jun 11, 2001; classic case of pineapple derangement syndrome

animist posted:

lmao

https://twitter.com/farbandish/status/1134485622352240640

rotor posted:

For real though anyone doing research in ML or machine vision should stop. The downsides of these technologies significantly outweigh the benefits.

# ? Jun 4, 2019 01:31

fart simpson: Jul 2, 2005; DEATH TO AMERICA

rotor posted:

For real though anyone doing research in ML or machine vision should stop. The downsides of these technologies significantly outweigh the benefits.

# ? Jun 4, 2019 01:39

Schadenboner: Aug 15, 2011; by Shine

rotor posted:

For real though anyone doing research in ML or machine vision should stop. The downsides of these technologies significantly outweigh the benefits.

# ? Jun 4, 2019 01:42

Perplx: Jun 26, 2004; Best viewed on Orgasma Plasma; Lipstick Apathy

but we can't have a machine vision gap with the Chinese

the first country to detect dissidence in a baby wins

# ? Jun 4, 2019 01:50

Bloody: Mar 3, 2013

rotor posted:

For real though anyone doing research in ML or machine vision should stop. The downsides of these technologies significantly outweigh the benefits.

# ? Jun 4, 2019 06:51

teppichporsche: May 11, 2019

The Management posted:

for those of you who haven�t worked with machine learning, here�s how it�s works:

* you take a bunch of pictures of shoes or horses or whatever the gently caress you want to recognize.
* then reduce those images to some ridiculous low resolution thing
* then you cram those images through some matrices until you get some matrix that happens to tell you horse or not horse.
* nobody knows why this matrix is horse, but it seems to work so you�re like, cool, I made a horse model.

now you take your horse model and set out to detect some horses. you take some pics and run them through your model and it�s all here a horse, there a horse, no horse there. but then there�s one picture of a horse and your model says no, not horse. and you have no idea why it�s not horse. the matrix is just a bunch of 8-bit values, it doesn�t even look like a horse. the answer, which you will never discover, is that the picture has a reflection of the sun in the water under the horse, which is not where the sun goes and your model couldn�t deal with that.

and then you realize that your model doesn�t understand poo poo, it hasn�t learned anything. it is incapable of learning, it�s a loving filter.

lmao yeah

# ? Jun 4, 2019 08:11

animist: Aug 28, 2018

does anybody know anything about tsetlin machines? are they more or less than the average amount of nonsense

# ? Jun 4, 2019 18:21

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

I know this

�Thou shalt not make a machine in the likeness of a human mind.�

# ? Jun 5, 2019 07:05

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

yeah this is the ml thread so i wouldn't worry about that being a problem

# ? Jun 6, 2019 00:11

Kudaros: Jun 23, 2006

I got my phd doing some cool neurophysiology stuff with ~one weird kernel trick~ and now do healthcare analytics for insurance nonsense for people who hired me on the basis of being a wizard, which I most assuredly never claimed.

Now I'm stuck repeating myself -- "If the predictors have no correlation with the response, no amount of ml in the world will help us" to be met with the response "Well this is the only data we have". Then I spend a week trying to demonstrate clearly that this won't in fact work. "What if you trim X?" "What if you put the answer in the features?"

Weeks later "Well lets look at this other data source" (that I could have used weeks earlier)

Review time: "you need to work faster"

Likewise,

Manager on phone with OCR vendor:

vendor: "what kind of accuracy do you need"
manager: "100%"

Welp.

# ? Jun 6, 2019 00:59

Share Bear: Apr 27, 2004

jupyter is really cool

# ? Jun 6, 2019 01:14

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

is "healthcare analytics for insurance nonsense" just another way of saying "killing people more profitably"?

# ? Jun 6, 2019 01:16

Share Bear: Apr 27, 2004

Deep Dish Fuckfest posted:

is "healthcare analytics for insurance nonsense" just another way of saying "killing people more profitably"?

yeah, specifically finding cases of diseases where the "long tail" of expenses are so high they kill out the average pooling

# ? Jun 6, 2019 01:18

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

i'd judge but i do lovely stuff for work without doing ml and without even the excuse of deece figgies plus i'm dead inside so whatever

# ? Jun 6, 2019 01:25

Kudaros: Jun 23, 2006

It's more like fraud and 'billing error' detection in my particular case. After my postdoc job fell through I applied for jobs letting people die more efficiently, amongst other things, but went with the only employer that called me back.

# ? Jun 6, 2019 01:41

flakeloaf: Feb 26, 2003; Still better than android clock

Deep Dish Fuckfest posted:

is "healthcare analytics for insurance nonsense" just another way of saying "killing people more profitably"?

numbing my pain with addiction
extending my "life" with a nurse

# ? Jun 6, 2019 13:41

Miss Mowcher: Jul 24, 2007; Ribbit

There�s lots of books of how to use ML to tackle loads of problems but is there a good one that talks about which kind of problems are adequate and which are not ?

# ? Jun 6, 2019 14:17

Cybernetic Vermin: Apr 18, 2005

ZZZorcerer posted:

There�s lots of books of how to use ML to tackle loads of problems but is there a good one that talks about which kind of problems are adequate and which are not ?

not sure about the word "adequate" here, but i think the answer is still: no, because we don't really know. there may be some literature on rules of thumb, but overall the most sophisticated current practice is throwing stuff at the wall and seeing what sticks.

# ? Jun 6, 2019 14:48

Bloody: Mar 3, 2013

an ethics course or two

# ? Jun 6, 2019 15:29

Miss Mowcher: Jul 24, 2007; Ribbit

I�ll try to get in the masters program at the CS department in my Uni later this year but another option that I was thinking was the Philosophy dept to try Law/Ethics in Computer/AI/ML :shobon:

# ? Jun 6, 2019 16:37

Share Bear: Apr 27, 2004

Bloody posted:

an ethics course or two

realtalk any decent ML book starts out with at least some of the reasoning and ethics behind using it, and then how it can be applicable

most of it is making sure your data ~~overfits~~ is good

# ? Jun 6, 2019 16:46

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

Bloody posted:

an ethics course or two

is ethics a growth market?

# ? Jun 6, 2019 17:20

Bloody: Mar 3, 2013

some flavor of engineering ethics course(s) should be mandatory for all stem degrees

# ? Jun 6, 2019 17:25

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

i doubt it'd have much on an effect. anyone who doesn't care about the impact of their work right now or cares but can't afford to do anything about it wouldn't be doing things differently because of a mandatory ethics 101 class in college

wouldn't do anything for sociopath executives either

# ? Jun 6, 2019 17:50

Kudaros: Jun 23, 2006

I went to grad school out of an intense curiosity and a desire to study the most pressing problems we face. Unfortunately those jobs are in short supply and, failing the profit motive, slow to grow or replenish.

Ethics on the individual level are well and good, but in an otherwise unorganized society I'm not sure what effect one should expect. I have high-demand skills and want to escape poverty forever, so I took a job. I'm looking for an out, but this is where I'm at now.

# ? Jun 6, 2019 17:56

Cybernetic Vermin: Apr 18, 2005

can't expect organized views and action from unknowing individuals though.

# ? Jun 6, 2019 17:57

Arcteryx Anarchist: Sep 15, 2007; Fun Shoe

all the ethics courses in the world won't matter when each individual is just some alienated contributor to a greater machine, and who can rationalize away their own involvement in anything horrible which may result from their work

# ? Jun 6, 2019 18:05

Silver Alicorn: Mar 30, 2008; 𝓪 𝓻𝓮𝓭 𝓹𝓪𝓷𝓭𝓪 𝓲𝓼 𝓪 𝓬𝓾𝓻𝓲𝓸𝓾𝓼 𝓼𝓸𝓻𝓽 𝓸𝓯 𝓬𝓻𝓮𝓪𝓽𝓾𝓻𝓮

they told me metal gear Rex was going to be used to *shoot down* nuclear missiles

# ? Jun 6, 2019 20:16

huhwhat: Apr 22, 2010; by sebmojo

Kudaros posted:

I went to grad school out of an intense curiosity and a desire to study the most pressing problems we face. Unfortunately those jobs are in short supply and, failing the profit motive, slow to grow or replenish.

Ethics on the individual level are well and good, but in an otherwise unorganized society I'm not sure what effect one should expect. I have high-demand skills and want to escape poverty forever, so I took a job. I'm looking for an out, but this is where I'm at now.

lancemantis posted:

all the ethics courses in the world won't matter when each individual is just some alienated contributor to a greater machine, and who can rationalize away their own involvement in anything horrible which may result from their work

# ? Jun 6, 2019 20:57

huhwhat: Apr 22, 2010; by sebmojo

mankind harnessing fire was a mistake

mankind was a mistake

# ? Jun 6, 2019 20:58

Lonely Wolf: Jan 20, 2003; Will hawk false idols for heaps and heaps of dough.

just make a NN that outputs "ethical" or "not ethical", problem solved

# ? Jun 6, 2019 21:19

animist: Aug 28, 2018

Lonely Wolf posted:

just make a NN that outputs "ethical" or "not ethical", problem solved

with a small perturbation of the input I can get it to say everything is ethical. so much more efficient than ethics boards!

# ? Jun 6, 2019 21:33

Lonely Wolf: Jan 20, 2003; Will hawk false idols for heaps and heaps of dough.

animist posted:

with a small perturbation of the input I can get it to say everything is ethical. so much more efficient than ethics boards!

ethical

# ? Jun 6, 2019 21:37

Bloody: Mar 3, 2013

Lonely Wolf posted:

just make a NN that outputs "ethical" or "not ethical", problem solved

public static String DetermineEthicalness(object input)
{
return "ethical";
}

# ? Jun 6, 2019 21:45

Lonely Wolf: Jan 20, 2003; Will hawk false idols for heaps and heaps of dough.

You stole my implementation and that's . . . ethical. DAMMIT

# ? Jun 6, 2019 21:55

huhwhat: Apr 22, 2010; by sebmojo

code:

if ethical:
	return not ethical
else:
	return ethical

# ? Jun 6, 2019 22:27

Adbot: ADBOT LOVES YOU

# ? May 11, 2024 21:33

Chopstick Dystopia: Jun 16, 2010; lowest high and highest low loser of: WEED WEE
k

huhwhat posted:

code:

if ethical:
	return not ethical
else:
	return ethical

woah I thought palantir's code was under NDA

# ? Jun 6, 2019 22:34

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > machine learning: this place is not a place of honor

«‹›24 »