Reminds me of an early application of AI where scientists were training an AI to tell the difference between a wolf and a dog. It got really good at it in the training data, but it wasn’t working correctly in actual application. So they got the AI to give them a heatmap of which pixels it was using more than any other to determine if a canine is a dog or a wolf and they discovered that the AI wasn’t even looking at the animal, it was looking at the surrounding environment. If there was snow on the ground, it said “wolf”, otherwise it said “dog”.
Hot dog. Not hot dog
That’s funny because if I was trying to tell the difference between a wolf and a dog I would look for ‘is it in the woods?’ and ‘how big is it relative to what’s around it?’.
What about telling the difference between a wolf and grandmother?
Look for a bonnet. Wolves don’t wear bonnets.
I can confirm this. I’m not a wolf expert, or even seen that many wolves really, but I have a dog and I don’t think she’d wear a bonnet.
Yeah, that’s a grandmother, so what?
While I believe that, it’s an issue with the training data, and not the hardest to resolve
So is the example with the dogs/wolves and the example in the OP.
As to how hard to resolve, the dog/wolves one might be quite difficult, but for the example in the OP, it wouldn’t be hard to feed in all images (during training) with randomly chosen backgrounds to remove the model’s ability to draw any conclusions based on background.
However this would probably unearth the next issue. The one where the human graders, who were probably used to create the original training dataset, have their own biases based on race, gender, appearance, etc. This doesn’t even necessarily mean that they were racist/sexist/etc, just that they struggle to detect certain emotions in certain groups of people. The model would then replicate those issues.
I bet ML would also think people with glasses are smarter or some dumb thing like that.
Old data adage. Garbage in, garbage out.
Actually, in this case the data sounds pretty clean.
Maybe not the hardest, but still challenging. Unknown biases in training data are a challenge in any experimental design. Opaque ML frequently makes them more challenging to discover.
The unknown biases issue has no real solution. In this same example if instead of something simple like snow in the background, it turned out that the photographs of wolves were taken using zoom lenses (since photogs don’t want to get near wild animals) while the dog photos were closeup and the ML was really just training to recognize subtle photographic artifacts caused by the zoom lenses, this would be extremely difficult to detect let alone prove.
Exactly.
The general approach is to use interpretable models where you can understand how the model works and what features it uses to discriminate, but that doesn’t work for all ML approaches (and even when it does our understanding is incomplete.)
Yes, “Bias Automation” is always an issue with the training data, and it’s always harder to resolve than anyone thinks.
Early chess engine that used AI, were trained by games of GMs, and the engine would go out of its way to sacrifice the queen, because when GMs do it, it’s comes with a victory.
Reg, why’d you just stab yourself in the shoulder?
Ah cmon, ain’t ya ever seen a movie?
Well of course I’ve seen a movie, but what the hell are ya doing?
Every time the guy stabs himself in a movie, it’s right before he kicks the piss outta the guy he’s fightin’!
Well that don’t… when that happens, the guys gotta plan Reg, what the hell’s your plan?
I dunno, but I’m gonna find out!
It’s not wrong
Why would you use AI for chess?
You don’t use it for the rule-set and allowable moves, but to score board positions.
For a chess computer calculating all possible moves until the end of the game is not possible in the given time, because the number of potential moves grows exponentially with each further move. So you need to look at a few, and try to reject bad ones early, so that you only calculate further along promising paths.
So you need to be able to say what is a better board position and what is a worse one. It’s complex to determine - in general - whether a position is better than another. Of course it is, otherwise everyone would just play the “good” positions, and chess would be boring like solved games e.g. Tic-Tac-Toe.
Now to have your chess computer estimate board positions you can construct tons of rules and heuristics with expert knowledge to hopefully assign sensible values to positions. People do this. But you can also hope that there is some machine learnable patterns in the data that you can discover by feeding historical games and the information on who won into an ML model. People do this too. I think both are fair approaches in this instance.
You can calculate all possible moves in milliseconds on any silicone these dsys
All possible moves one step from a given position sure.
But if you then take all possible resulting positions and calculate all moves from there, and then take all possible resulting positions after that second move and calculate all possible third moves from there, and so on, then the possibilities explode so much in number that you can’t calculate them anymore. That’s the exponential part I was refering to.
You can try and estimate them roughly, let’s say you’re somewhere in the middle of the game, there are 12 units of each side still alive. About half are pawns so we take 1.2 possible moves for them, for the others, well let’s say around 8, thats a bit much for horses and the king on average, but probably a bit low for other units. So 6 times 8 and 6 times 1.2, lets call it 55 possibilities. So the first move there are 55 possible positions, for the second you have to consider all of them and their new possibilitues so there are 55 times 55 or 3025, for the third thats 166375, then 9.15 million, 500 million, 27.6 billion, 1.5 trillion etc. That last one was only 7 moves in the future. Most games won’t be finished by then from a given position, so you either need a scoring function or you’re running out of time.
Yep, those are the moves that can all be easily calculated very quickly on modern hardware
There are more possible chess moves (estimated at 10^120 for an average game) than there are atoms in the observable universe (estimated at 10^80). That is to say the number of possible chess moves has 40 more zeros on the end than the number of atoms in the observable universe.
Can you point to some souce showing how modern hardware can work these out easily?
No, they can’t
That shit works IRL too. Why do you think therapy practices often have themselves positioned in front of a wall of books? Not that it’s a bad thing; it’s good for outcomes to believe your therapist is competent and well educated.
Maybe true but your comment is humanizing “dumb” AI.
Someone should build a little AI app that scrapes a job listing, then takes a resume and rewrites it in subtle ways to perfectly match the job description.
Let your AI duke it out with their AI.
Pretty sure I saw that somewhere actually.
Edit: Top 5 List - https://www.tomsguide.com/features/i-tested-5-ai-resume-builders-to-help-get-a-job-heres-how-they-fared
I’m looking for a job now, and this is very useful to me. Thank you.
deleted by creator
That exists
Do you really want to work for a company that allows their HR department to abuse AI as a tool?
Explain how anyone has a choice, especially in the United States
Yes? They’ve got a million bazillion applicants too; this is a huge market failure all around.
When I got out of the military, my outprocessing included a briefing about how to get interviews with federal organizations. One thing they taught us was that you can copy the job description, paste it into your resume, and set the font to white. The automated systems at USA Jobs would register a match to the job description and rate you as a better candidate and the human screeners were so overworked that they would just go with what the computer says without checking.
I wonder if this still works lol that’s brilliant.
Some places, yeah. But some of the screening tools have grown more sophisticated now that the information about copying and pasting the listing has made its way to the public.
Well, at least they didn’t spend a lot of money on testing it…
I don’t understand why anyone writing, reading or commenting on this think a bookshelf would not change the outcome? Like what do you people think these ml models are, human brains? Are we still not below even the first layer of understanding?
The problem is the hysteria behind it, leading people to confuse good sounding information with good information. At least when people generally produce information they tend to make an effort to get it right. Machine learning is just an uncaring bullshitting machine, that is rewarded on the basis of the ability to fool people (turns out the Turing test was a crappy benchmark for practice-ready AI besides writing poems), and VC money hasn’t reached the “find out” phase of that looming lesson, when we all just get collectively exhausted by how underwhelming the AI fad is.
Yeah, the hysteria is definitely the problem. Can’t say that I agree that the technology is underwhelming, though. It can generate, practically anything fast and with guidance and it’s just interesting that nobody really understands how. It’s a paradigm shift for creative work. Producing music or art will continue to change a lot from this. Using the technology to analyse personalities during job interviews is so fundamentally idiotic, because a generative system is a brainstorming tool, not analytical nor accurate. And just so wrong that it feels like it’s actually the work of someone malicious.
I wonder if it’s actually interpreting the bookshelf or if having such a busy background is taking a toll on the compression. That would alter the details on the person’s face
Good and interesting question. I bet you could test it by using static (high entropy) as a background vs the control plain color.
I would be interested to see what happens if you lighten up her skin color a bit…
Conventionally attractive white people, stealing all your jobs!
Go full albinism
During the AI goldrush you can make your fortune selling bookshelves.
Selling bookshelf large poster, or just jpgs
Having a bookshelf poster behind you is actually a hilarious workaround.
Bookshelf NFTs? Only possible to buy with crypto?
Answering the question in the image: machine learning arose from the industrial control world. The idea was to teach a machine how to detect defects in supposedly identical objects out of a manufacturing line, most often with “machine vision” (ie. a camera). Applying it to humans was asinine.
I know right? I have seen seen vision systems do some impressive things, but they are carefully calibrated to work in a specific way under certain conditions. Some of the ones my company works with get fed CAD in real time so the robot knows what to look for.
“Bias automation” is kind of an accurate description for how our brains learn things too.
The base assumption is that you can tell anything reliable at all about a person from their body language, speech patterns, or appearance. So many people think they have an intuition for such things but pretty much every study of such things comes to the same conclusion: You can’t.
The reason why it doesn’t work is because the world is full of a diverse set of cultures, genetics, and subtle medical conditions. You may be able to attain something like 60% accuracy for certain personality traits from an interview if the person being interviewed was born and raised in the same type of environment/culture (and is the same sex) as you. Anything else is pretty much a guarantee that you’re going to get it wrong.
That’s why you should only ask interviewees empirical questions that can identify whether or not they have the requisite knowledge to do the job. For example, if you’re hiring an electrical engineer ask them how they would lay out a circuit board. Or if hiring a sales person ask them questions about how they would try to sell your specific product. Or if you’re hiring a union-busting expert person ask them how they sleep at night.
I’ve just started doing practical interviews. I basically get really young people with little overall experience and I just want to know if they can do common technical tasks.
So one question is to literally have them explain how to tighten a bolt. One person failed.
To be fair, that’s a very open ended question. I mean, what kind of bolt are we talking about? A standard lag bolt? If so you don’t tighten it! That’d be a trick question! You tighten the nut. Same thing applies with car wheel bolts. Tricky tricky!
Is it a hex bolt that also has a cross head? How tight are we talking?
I’m just going to assume bolts of lightning and Usain Bolt are off the table.
I’m just going to assume bolts of lightning and Usain Bolt are off the table.
The only thing I know about the procedure for tightening Usain Bolt is that I am not part of performing it.
I did actually make the mistake of asking just “which way do you turn a screw” once and the person had the sense to ask “to tighten or loosen it?”
Would you have accepted “righty tighty lefty loosely”?
Yeah but if they don’t show which is which I ask them to show too.
Almost everyone gets screw turning right, it just weeds out a few people who say the right things in emails.
Not really in a bolt tightenning domain, but I have done technical interviews for a lot of devs including junior ones, and them asking all those questions about the task is something I would consider a very good thing.
At least in my domain the first step of doing a good job is figuring out exactly what needs to be done and in what conditions, so somebody who claims to have some experience who when faced with a somewhat open ended question like this just jumps into the How without first trying to figure out the details of the What is actually a bad sign (or they might just be nervous, so this by itself is not an absolute pass or fail thing).
But all the other questions are to find out if they are a good fit for the office culture.
You know, if they are also white middle class dude bros.
That’s why you should only ask interviewees empirical questions that can identify whether or not they have the requisite knowledge to do the job.
Hol up. ThAt sOuNds LiKe RaCisM!
I really hate that we are calling this wave of technology “AI”, because it isn’t. It is “Machine Learning” sure, but it is just brute force pattern recognition v2.0.
The desired outcomes you define and then the data you train it on both have a LOT of built-in biases.
It’s a cool technology I guess, but it’s being misused across the board. It is being overused and misused by every company with FOMO. Hoping to get some profit edge on the competition. How about we have AI replace the bullshit CEO and VP positions instead of trying to replace fast food drive through workers and Internet content.
I guess that’s nothing new for humans… One human invents the spear for fishing and the rest use them to hit each other over the head.
I agree with most of your points, but i don’t entirely like the “this is not intelligence” line of thought. We don’t even know yet how to define intelligence, and pattern recognition sounds a LOT like what our brains do. The hype is of course ridiculous, and the ways it’s being used is just stupid, but i do think pattern recognition could be a solid basis for whatever we end up considering intelligence.
All you are saying is, is that intelligence isnt as smart as we think, that human intelligence is actually pretty dumb. That doesnt change anything about the current situation even if thats true though.
So we all agree its actually human level intelligence now, what then? Can we stop developing it and do something else now?
Naw, humanity needs to waste more electricity.
Lol, wtf XD
“That doesnt change anything about the current situation even if thats true though.”
Yeah, i assumed me writing “I agree with most of your points” conveyed that. Do you always imagine random things to attack instead of just reading what people actually write?
wtf O_o
I just don’t like people being like “but it’s not real intelligence” while we don’t even know what intelligence is, and we’re thus avoiding the one part of this stupid hype that could be interesting:: philosophical questions about our own intelligence/humanity/…
Do you always argue the most boring parts of any issue as a rule? You win the argument, congratulations I hope it changes the world.
Do you always change the topic to try and “win” online discussions? And act so unnecessarily hostile for no reason at all to people who want to have interesting online discussions?
I find the topic of whether it’s intelligence the most interesting part of this. It raises a lot of questions. That the current hype is ridiculous that a lot of the energy expended on it is a complete waste, and that most of the ways AI is used is beyond stupid isn’t even worth talking about, that’s just plain obvious.
I think you might be projecting your own hostility there.
And you are just hijacking the AI conversation to argue about what the word intelligence means. I thought you wanted to talk about AI thats why I originally replied. When I realized you just wanted to apparently correct the public about their use of your favorite word, I decided this wasnt worth it.
Its sort of like you went to a bowling forum, and were very excited to discuss the mineral contents of the oil on the bowling lane, but got upset when noone cared to discuss with you because they just want to talk about bowling.
Dude, just stop. You’re looking for things that aren’t there. Period.
I find it an interesting question whether it’s intelligent or not, and find it sad people throw out that question together with the rest of the hype. It’s not “my favourite word”, and i’m not projecting my hostility. You just can’t seem to handle someone bringing any bit of nuance to a discussion…
And me projecting hostility XD. yeahhhh… introspection isn’t one of your gifts it seems XD. I’m the one being hostile XD. roflmao XD. I mean just here “your favorite word”… wtf dude, exaggerate much to make yet another pointless jab at me?? I’m not allowed to find this an interesting question without you painting me as someone who fixates on that one thing in the world and makes it sound as if my world revolves around “AI IS INTELLIGENT!!!”… I’m not even convinced it is, but i find it a mighty interesting question that requires more thought than it’s getting.
Sorry for trying to argue something i find interesting on lemmy. I’ll just shut up next time and not try to bring up points you might not find interesting since you seem to take that as a personal offence, while you could have just shut up and let the adults have a nice conversation on the one interesting part of this hype.
Maybe it is human-like intelligence. It’s dumb as shit, but have you met people?
LMAO
But yeah, I guess at its core, human intelligence and machine intelligence are both just pattern recognition, but I guess my point is that calling it “AI” gives people this false sense that it is something it is not. AI has been a thing in Sci-fi for so long that we all think of Data from Star Trek or C-3PO from Star Wars and similar. When in reality it is more akin to a robot arm in a factory doing the same task really fast and really precisely, but it isn’t some adaptable all-purpose thing yet.
That for sure is a problem with all modern bullshit technologies they want to hype in order to get people to use/buy it.
Look at smart tv’s… everyone assumes they’re awesome since they’re smart tv’s, that’s of course better than a regular tv. They’ll of course never mention that this just means that it’s a tv with a 100$ android box embedded that they’ll abuse to try to serve you extra ads, that they’ll not bother to update so your tv becomes obsolete in a couple of years, and that you can achieve the same thing by just buying the android box sepearately and connect that to a regular tv, which won’t make your entire tv become obsolete when the cheap android box doesn’t get updated anymore…
So yeah, i can imagine you have an issue with it being marked as (competent) AI.
Yeah for real.
Smart TVs, Subscription services, etc.
It’s all just capitalism doing its thing, everyone racing to sell sell sell.
Pattern recognition is one thing that our brains do, it is a very long way away from the only thing our brains do.
[citation needed]
But yeah, that’s the kind of discussion i’d love to see in more depth :). When would an AI be considered intelligent? It used to be passing the turing test, but now that’s being achieved the goalposts are moving, and that’s maybe for a good reason, but what will be the actual measure :).
That reminds me of the time, quite a few years ago, Amazon tried to automate resume screening. They trained a machine learning model with anonymized resumes and whether the candidate was hired. Then they looked at what the AI was looking at. The model had trained itself on how to reject women.
Another similar “shortcut” I’ve heard about was that a system that analyzed job performance determined that the two key factors were being named “Jared” and playing lacrosse in high school.
And, these are the easy-to-figure-out ones we know about.
If the bias is more complicated, it might never be spotted.
“Machine learning” is perfectly cromulent. The bias is what it learned, because that’s what it was taught. (Not intentionally, I don’t think. It’s just hard to get this stuff right sometimes.)
Job interviews are all bias regardless of whether they’re automated 😅
I’m really good at my job.
But that’s not why I got my job, it’s just a coincidence.
I got my job because I’m pretty good at interviews.
It is bias laundering though. They hide behind an “objective” algorithm, which was trained on a huge dataset of past
biasedhiring decisions.Garbage in garbage out
Garbage all around
The idea of AI automated job interviews sickens me. How little of a fuck do you have to give about applicants that you can’t even be bothered to have even a single person interview them??
I dunno, but if your boss chain contains a machine (literally Amazon warehouse), does it matter?
But god forbid the applicant didn’t spend hours researching every little detail about a company, writing a perfect letter with information that could have just been bullet points and being able to explain exactly why they absolutely love the company and why it’s been their dream to work there since they were a child. Or even worse: Use AI to write the application.
Exactly!
Applicants are expected to dedicated hours of their time to writing their application and performing background research - both of which are becoming increasingly more tedious over time - so the least a company could bloody do is show some basic respect by paying an actual human being to come interview you!
We should build an AI that automates researching about a company for applicants
The real “No U” of AI…
Cover letters fucking make me hateful. I love generating AI cover letters and sending them. Fuck your cover letters in a market where you need to send 100 applications to get 10 bites
That’s more like an excuse to keep those stupid 5, 6, and even more interview round processes. Basically making you work an entire week for free in exchange of a chance of getting an offer. Make the first or second rounds with AI and only bother after that.
“oooo books he must be really smart”
