> One of our conclusions is that not everything on Twitter is a good candidate for an algorithm, and in this case, how to crop an image is a decision best made by people.
This seems like it should have been a foregone conclusion. What was the driving force in the first place to think cropping images with an AI model was desirable? Seems like ML was a solution looking for a problem here, and I'm glad they've realised that.
Right but... we've been cropping images in web applications since... y'know, pretty much ever. Using ML to do this was always pretty ridiculous overkill. Give the users an image cropper, and be done with it.
I can't see why this is overkill. You're eliminating a step from the image posting process, and making it so users don't have to crop an image twice (once for the full image, and a second time for the preview). That makes sense when you're writing a CMS or blogging platform like Wordpress, but for Twitter it adds some friction.
So, previously, the preview was just cropped in the center. But this made some images look funny, since people's faces would get chopped off.
Coming up with a workable solution to this with ML is not especially hard. You can get things like face detection off the shelf, maybe just tell your autocropper, "crop closer to the face" and have a demo within a couple days (and then much more effort to productionize it). From there, you can start introducing ML models to improve on your basic face detection. (I'm not counting face detection as ML.)
This is not a case where some massive ML model is being brought in to save two seconds of your time. This is a very natural and obvious application of ML, at a company which already does ML at scale, in a way that sounds like it has a good chance at improving the appearance of the site without introducing additional friction.
Instragram gets around this by encouraging everyone to take square photos.
I don’t think anyone is saying “I will always prefer to crop every photo and everyone else should too”. I think the point is closer to, if I may borrow a Simpsons line, “I liked your half-assed underparenting a lot more than your half-assed overparenting”. It’s actually impressive that Twitter didn’t object “but I was using my whole ass”, which is basically their default trope when they address user complaints.
All those examples show large improvement. Of course they might cherrypick images with large improvement for their blog advertising the feature. But still, it illustrates why people would think it's a good idea.
Of course they don't seem to consider the idea of not cropping at all.
I'm more forgiving about corporate jargon than most. A lot of it really does help optimize communication for the situations you encounter in corporate work.
But "learnings" is literally, exactly, just a synonym for "lessons." Can we not?
I disagree, "Sharing lessons..." would mean "here is an educational resource that we have created, as teachers, for an audience of students". I think "Lessons learned..." is closer to what you mean to suggest, and "Learnings..." is more concise (this is from Twitter, after all).
It's a neologism, or possibly the resurrection of a long unused form - I don't know exactly how we came about it but I agree completely that one of the meanings of "lesson" is "a thing which has been learned".
In my experience, there's a tendency toward folksiness in certain varieties of corpspeak that causes rejection of "formal-sounding" terms and repurposing of "plainer" forms to create new words, hence lessons = learnings, protégé = mentee, and so on.
I don't think the fact that some words can be formed that way is a good guide to whether other formations are words or not.
Do you think "cookings" is a word? It would be "things that have been cooked" I suppose, on that basis. I would say if it's a word it's one I've never heard before.
On the other hand, "cuttings" is clearly a word in common usage in horticulture.
My point is more that it sits in roughly the same semantic space as ‘lessons’ and ‘learnings’, yet there seems to be room for nuanced differences of meaning. Yet I don’t see people object to ‘teachings’ on the basis that we already have the word ‘lessons’.
> I don't know why people object to the word 'learnings'.
It bugs me because it emerged out of tech executive culture and caught on in large part because others used it as an aspirational signal to sound like they are members of that group.
I don't really care, but since you asked: "learnings" grates my ears because it sounds wrong. I never used to hear people say it, and as GP points out, we already had the perfectly good word "lessons".
"Learnings" to me sounds like a mistake a non-native English speaker would make, perhaps by incorrectly extrapolating from the word "teachings". It's a forgivable mistake, but I don't see why we need to adopt it into the native vernacular.
(Yeah, yeah, I know, language evolves, prescriptive versus descriptive etc.. In a world where "literally" no longer means literally, I'm not going to waste much time getting angry about "learnings".)
Bias aside, the saliency algorithm doesn't work well either. This twitter feed (SFW) https://twitter.com/punhubonline often shows the punchlines in the preview, spoiling the joke.
I agree that a manual option should have always been available. I'm just being pedantic: the context here is a "saliency algorithm." In the context of computer vision, the term saliency has a specific meaning. It's about visual perception of images, which doesn't generally include understanding letters, words, sentences, or humor.
Sure, if it's designed for cropping cartoon images to show the first panel. This algorithm is designed to work for any image shared on twitter. What should it do for cartoons where the last panel is a much more interesting visual image? Single-panel cartoons that include both the setup and the punchline in one frame (https://twitter.com/PunHubOnline/status/1392509471482683399/...)? How about before-and-after comparison images?
So, I can choose to see only un-cropped images on my TL, and the author can see a preview of the algorithm's crop before they tweet -- but a glaring omission is simply exposing a crop tool to the author. The model works by choosing a point on which to center the crop. Why can't you give user's a UI to do the same? "Tap a focal point in the image, or let our robot decide!"
The blog post mentions several times how ML might not be the right choice for cropping; but their conclusion was...to keep using ML for cropping. I hope someone got a nice bonus for building the model!
I can't really see any down side, besides maybe a little bit of developer time, to allowing users to see a preview of the crop and optionally override. It's done all the time in other places.
It's probably a bit harder at Twitter's unique scale. They have an incredibly high throughput of new posts and a large portion of these posts include between 1-4 images that need cropping.
Isn't this just a UI change and wiring that up to the backend? The cropping happens either way, it might actually be faster considering that if the user providers the crop info, Twitter does't have to burn CPU cycles on figuring out the "optimal" crop dimensions.
Well, no it wasn't sarcasm... if Twitter can offload image attach and crop entirely onto the client, I figure they would be doing that. Certainly their frontend engineers understand how to. But they must have some reasons for involving the server.
Because they do!
If you go attach an image to a new tweet, you will see it gets uploaded right then and there, whether or not you ever hit post. Who knows what happens from there. Allowing clients to demand more crop work, at will, on top of the one that is automatic seems like it would have some implications at twitter's scale. At the very least, it would probably undo some optimizations they've made around image handling.
> but their conclusion was...to keep using ML for cropping
My takeaway from the article was that their conclusion was to remove cropping from the product, starting incrementally on iOS. (I got cropping removed on Android as well recently). That seems like the opposite of "keep using ML for cropping"?
(Note the thread displays differently now because Twitter have changed their cropping algorithm)
Originally @colinmadland was trying to post examples of how Zoom virtual background had removed his black colleagues head, however when he posted the side-by-side images (with heads) on Twitter, twitter always cropped out his colleague and just showed him, even if he horizontally swapped the image. So, while trying to talk about an apparently racist algorithm in Zoom, he was scuppered by an apparently racist algorithim in Twitter.
The web version shows Mitch, but the app shows a blank white (which is at the center of the image, meaning it didn’t try to crop to one of the faces). I’m on iOS.
That example is from last September, so it doesn't say anything on if it is improved or not. They probably generate the cropping once, on posting the tweet.
I find that calling it a `racist algorithm` doesn't really do it any good unless the behaviour was intentional. This is a case of poor training data the same as google image classification messing up with tags.
Plenty of racism in humans isn't malicious, either, but is just a byproduct of bad training data. The outcome is bad regardless of what was intended, and it's the outcome that matters.
Matters how? If someone has the best of intentions but ends up creating a bad outcome, it’s still more important to fix that than to change the opinion of, say, a closeted bigot who has no effect on anyone else. (Yes, real life has more shades of gray than that, and both things are important in practice, because bad intentions don’t tend to lead to good outcomes while good intentions definitely can lead to bad outcomes.)
As an aside, “I think plenty of people would say X” is not a very good way to phrase an opinion. It’s ok to say you would say X and argue for it, rather than ascribe the opinion to some undefined group of other people.
Racism as a concept has evolved in meaning. It used to only include the most severe intentional cases of bigoted behavior, whereas now it also includes less obvious biases that lead to preventable but not necessarily intentional instances of everyday prejudice and bigotry.
I am for one happy we have unneutered the word from having to reach a bar so high, it wouldn't apply to most bigotry, but it is also unfortunate for people who have not caught on and believe calling a thing racist is a damning statement of evil intent, but it really is not anymore. Or those that insist on meaning of words remaining static forever.
Let's say your company decides to use AI to assist in hiring. It turns out the algorithm used is biased when it comes to candidates' race. If there is a disparate impact[1] on protected classes in hiring that's unrelated to job performance, intentions don't matter in the eyes of labor law, what matters are the effects.
ok we need pictures of human faces, luckily I've got all these white people here!
on edit: it was racist in result, in that it empowers a racist system, it was not racist in intention - as in the people gathering the training data probably didn't say hey how can we empower a racist system with this?
The Zoom example is a racist algorithm. It was built using a against a data set that produced different results for different skin colours.
The Twitter example was not a racist algorithm. It would consistently pick one head over the other, but it had nothing to do with the skin colour. It might preference the black head for some pairs, and the white head for other pairs.
In the second example people anthropomorphised the algorithm. They assumed that any example of a preference for an images was due to a racial bias. It was easy to keep feeding it images to get to an input that confirmed this assumption.
Image cropping algorithms are hard. When we made our first one for reddit, it used this algorithm:
Find the larger dimension of the image. Remove either the first or last row/column of pixels, based on which had less entropy. Keep repeating until the image was a square.
The most notable "bias" of this algorithm was the male gaze problem identified in the article. Women's breasts tended to have more entropy than their face, so the algorithm focused on that since it was optimized for entropy. To solve the problem, we added software that allowed the user to choose their thumbnail, but not a lot of users used it or even realized they could.
I assume they've since upgraded it to use more AI with actual face detection and so on, but at the time, doing face detection on every image was computational infeasible.
I see that you didn't get the memo. If an algorithm or a mathematical definition is made by a white man, it's by definition racist and sexist because it was made by a white man. Entropy was defined by Claude Shannon who was a white man, so entropy is racist and sexist, because it absorbed all of Shannon's biases.
Sadly he’s too angry to make the joke funny. As always there probably is a grain of truth to the idea that postmodern feminist critiques occasionally disappear up their own fundamental, and so there’s a joke to be made here. But this isn’t it!
Could have something to do with literally all major sources of information in US showing ideologically-charged garbage in front of people who want to focus on other things in life.
Entropy usually boils down to "sum p * log(p) for all p (and possibly normalize)", assuming you have discrete probabilities p. It is not related to variance or mean.
Which is not really image entropy at all, as it totally neglects spatial structure. You could sort an image and get the same entropy using the histogram approach.
I don't think the claim is that the behaviour is caused by "male gaze", but rather that the outcome of always focusing the cropping around any visible cleavage is functionally identical.
Whether or not it's unsupervised, whether or not it's sexist, it seems that a thumbnail focusing on a person's face rather than their breasts is typically going to be more desirable. Depending on context, of course.
> why get so moralistic and high-horse about just one?
Not doing so; just observing the facts.
If a proposed 'unsupervised' algorithm of this simplicity highlighted women's faces perfectly, but zoomed in on men's receding hairlines, it wouldn't have made it past the drawing board. Indeed, it's reasonable to believe that nobody would have noticed that it consistently worked for women. We certainly wouldn't know that this algorithm existed or be talking about it here.
We observe a bias in what is considered important to check before shipping.
That's an AWFUL lot of assumptions, there. You've constructed an entire complex narrative around something where all we know is that it's very simplistic.
I don't think this is a complex narrative. It's the reality of development:
The simplest case is just to pick the most central square, and then you could probably improve that by picking a standard square according the rule of thirds. Those are the naive algorithms - this choice of alternative algorithm is deliberate and isn't as naive or simplistic as you're claiming.
The algorithm is only considered useful because it appears to do better than that, on whatever examples that the developers tried (i.e. there was a business case for using it), and against other possible code.
Including, likely, pictures of their own selves. That's certainly what I would test it on, until it vaguely worked.
What 'awful lot of assumptions' do you think I am making? I don't imagine we are in disagreement about this.
Blasting out an ML algorithm in an afternoon that causes millions of people to see wildly different representations of people depending on their race or gender seems, maybe, like a bad thing you wouldn’t want to defend?
Sure, more testing would be better, but nobody cares how many landscape shots it messed up, right?
Years ago reddit was, IIRC, not very staffed at all compared to their traffic. It's a pretty privileged take to say they should have done expensive QA entirely around your particular things that you care about.
But that’s the problem. The entire premise of these algorithms is based around what you (or the developers producing it) care about. It’s the common thread between image crops preferring white faces and women’s breasts, and automated cars preferring dead black pedestrians over vehicle collisions.
If you don’t have the capacity to use new technologies without increasing harm, maybe you don’t have the capacity to use them.
No, I'm positing the premise that it was shipped with an absolute minimum of QA that didn't approach the level of trying to build 'inclusive' reference sets, on the cheap. It wasn't about caring in terms of priority of what was tested, it was about NOT caring that much in any direction and shipping it.
And it was a naive image cropping algorithm, years ago, and not making use of any sophisticated 'new technologies'. The beauty of the algorithm is that it was a simple function that could have been written in 1975 and required no training, deep learning or any of that. If you want to talk self-driving cars, you've got a much more relevant measure of harm and I'm right there with you.
As it is, I'd say there's a disconnect between where you and those years-ago shoestring developers stand on Maslow's hierarchy of needs. They were being scrappy with limited resources, and you're mad at them for not having an amount of QA that would have seemed unbelievable to them under their resource constraints.
> No, I'm positing the premise that it was shipped with an absolute minimum of QA that didn't approach the level of trying to build 'inclusive' reference sets, on the cheap. It wasn't about caring in terms of priority of what was tested, it was about NOT caring that much in any direction and shipping it.
Right. I’m saying that not caring is the problem. We mostly agree on the facts. I’m objecting to not caring as an acceptable position.
We’re not talking about shoestring developers. We’re talking about platforms that millions of people including heads of state use to affect billions of people’s lives.
> We’re not talking about shoestring developers. We’re talking about platforms that millions of people including heads of state use to affect billions of people’s lives.
Jedberg was >5 years ago, maybe closer to 10 for this algorithm. They were relatively shoestring compared to now, and especially compared to current-day FAANG.
Well I don't see any sense continuing this discussion, since you just keep refusing to engage with the harm done by doing something without caring about its impact.
It's an incredibly privileged definition of 'harm done'. They had less budget and manpower than you wish they had, and their shitty off-brand cropping algo occasionally had too much titty. If that's the level of harm that registers with you, you're clearly doing very well.
I’m gonna regret responding after saying I was removing myself but I’m hoping this will help you or at least other people reading understand what the harm is.
I’m not objecting to an algorithm that exposes me to more cleavage. I couldn’t care less. I’m objecting to an algorithm that exposes millions of men to a subtle but routine equation of men with faces, and women with breasts. It’s not a privilege for women to expect not to be casually sexualized because someone thought their algorithm was helpful but didn’t bother to find out it wasn’t.
Likewise the impact of racial bias. I’m not offended by seeing white faces. But it’s not a privileged position for POC to object to being basically erased from images because whoever developed the algorithm didn’t bother acknowledging more than one skin tone exists.
These things have real impact, not just on how it affects the people who are directly underrepresented. It also affects how the people who are overrepresented perceive and ultimately act toward them.
This sort of discussion makes me think of a book I once read by someone who was raised in a fundamentalist Christian sect, maybe Pentecostalism? Anyway, they described how, in their upbringing, the devil and his temptations was so omnipresent, that it was almost as if he was more powerful than God. Whenever you took your mind off the dangers of sin, the devil would come in and take control, one was continually warned. Eventually, this created a kind of cognitive dissonance, because God was remote but the devil was always by one's side, whereby they ended up losing their faith and becoming atheistic.
If we are aware of a negative principle that infests and invades every part of life, though we do our utmost to repel it and mitigate it and remediate it, how can we maintain the conviction that it is not the natural order of things, but is a product of solvable human faults?
I'm not denying the importance of anti-racism or of examples of bias as mentioned here; my point is that when faced with an all pervasive adversary, whether a personified devil or something more abstract, human minds tend to find relief in submitting to the seemingly inevitable via some rationalization.
Nobody is trying to assign blame to the person who wrote the algorithm, it's just being pointed out that the the output it produces is sub optimal in this specific way.
Breasts shouldn't have more entropy than face. Perhaps the reason is due to the breasts being in the middle of the picture, so the face gets being compared to bottom rows more frequently?
Why not? Shirts might have flashy patterns, differently colored fabrics, alternating skin and shirt. On a row-per-row basis I can see the chest area being more entropic than a face with an even skin tone.
edit: I googled "woman" and selected random pictures which showed the whole upper body, entropy summed over each row to the right: https://imgur.com/a/oVB57gu
Seconded. I'm picturing my wife's choices in clothing--I would say often the highest entropy will be her top. And depending on the resolution and how good the entropy measurement is I can easily see her pants coming in second. (She tends to favor repeating patterns.)
(She has no work dress code she needs to conform to. Rarely does she wear a top as plain as any of your examples.)
And even just trying to access a post/thread on mobile. I already clicked the link, then I have to click it once more to say “yes I want to do the browser thing I explicitly did”, then another time still to actually show more than half a screen of content.
"We began testing a new way to display standard aspect ratio photos... without the saliency algorithm crop. The goal of this was to give people more control over how their images appear while also improving the experience of people seeing the images in their timeline. After getting positive feedback on this experience, we launched this feature to everyone."
So the solution all along was to give users the ability to crop their own photos. Why wasn't this the original way of doing things?
Instead of forcing a complicated algorithm into the Twitter experience, it seems to me that the solution all along was just to let users do what they do best-- make tweets for themselves. This incident strikes me as a major failing of AI: We are so eager to shoehorn AI/ML into our products that we lose sight of what actually makes users happy.
> Why wasn't this the original way of doing things?
Someone wanted to do a feature so they could get promoted. Probably with some mumbo jumbo about how it reduces the number of clicks to create a tweet and thus increases revenue.
What’s really remarkable is that giving users the ability to manually crop would be an amazing way to gather data on optimal cropping, which they could have used to train their model down the road. I can only imagine how much more time and money went into gathering eye tracking data.
If you were trying for real bias in your cropping algorithm, I would suspect training it on what the average, unconsciously-biased user thinks is the best crop nearly guarantees it.
/rant but I feel like talking about percentage points of difference is always hard for humans. For example:
> In comparisons of men and women, there was an 8% difference from demographic parity in favor of women.
would have been clearer (and more correct) as "an 8 percentage-point difference from demographic parity". That 8 pp difference though is a 16% "relative" difference (58/50), or more starkly "The algorithm chose the woman almost 40% more often" (58/42 => 1.38). That said, the diagram in the post [1] is much easier for humans to parse and say "wow, that looks pretty far off!".
tl;dr: A number like 8% sounds like "no big deal", but 8 percentage points (on each side) is a big deal!
The article says the algorithm was trained on eye tracking data, and if the predominance of data came from men then referring to the male gaze as a possible source of unintended bias seems valid.
> It's disappointing that Twitter is trying to deflect blame for their failures onto "men".
They...aren’t. Leaving aside the question fo whether identifying the “male gaze” as a cobtributor to problems with the algorithm’s output would be blame passing when the entire context is taking responsibility for and addressing problems resulting from the choice of algorithm (hint: no, it wouldn’t), they initially identified that as a potential thing the saliency algorithm might* be biased for, constructed a test, and determined it was not a factor, so they are saying “male gaze” wasn’t a problem.
Kneejerk reactions to the use of the phrase without considering anything about what it is being used to say aren't helpful.
> In comparisons of black and white individuals, there was a 4% difference from demographic parity in favor of white individuals.
It's hard to believe that the bias was only 4% - there were a lot of people testing with images that they sourced themselves, and the preference for white people seemed much closer to 80-20.
The paper authors mention that their training data is from Wikidata (pictures of celebrities). I wonder if the types of photos in that dataset are meaningfully representative of the kinds of photos that people usually post to Twitter.
> It's hard to believe that the bias was only 4% - there were a lot of people testing with images that they sourced themselves, and the preference for white people seemed much closer to 80-20.
It's very easy to believe the bias was near-zero given you are citing highly motivated people on Twitter cherrypicking from thousands of examples and a little baffling you find that to be more credible than controlled systematic experimentation; note, for example, the extremely striking fact that the fuss completely missed the other bias they found which was several times larger - showing how totally useless people on social media are for testing these things and how they can conjure up "80-20" biases which don't exist.
> highly motivated people on Twitter cherrypicking
One of the reference threads that identifies the issue happened on it by accident highlighting a surprising experience in another product (Zoom). Believe it or not, people who care about this stuff are not looking for stuff to complain about, we’re tired and overwhelmed. And I would hope that people who, upon discovering a vulnerability find and catalogue the ways it can be exploited, would be celebrated here.
I admit that it wasn't especially rigorous testing, but I personally tested this along with other people I knew. I used real photos from my camera and my wife's (we are of different races), featuring photos of ourselves, friends, and family.
I of course hope that the systems I use aren't racist against my loved ones. I am motivated to confirm whether or not they are, but I didn't go on to parlay my findings into an essay for clout. I gained nothing from doing this, except the knowledge that Twitter was suckier than I knew.
Possibly a political question but why is the word "equitably" more popular now than "equally"? I'm not sure when I first noticed this but it seems pretty recent that "equity" became more used than "equality" when referring to diversity and inclusivity
Because sometimes addressing inequality requires addressing situations unequally. As a (hopefully) completely uncontroversial example: if you were a governor or president during Civil Rights era desegregation, evaluating whether to deploy National Guard soldiers to enforce desegregation of schools, it would be absurd to apply that equally to schools which are segregated and schools which are integrated.
This is a very cynical take (I'm cranky from my second vaccine dose), but:
Imagine how much work, how much energy and effort, went into building an ML-based image cropping feature, just because an up-and-coming Designer decreed that a certain specific crop ratio was the most aesthetically pleasing to the human eye...
...so that years later, after countless hours of additional user research, they would just remove the feature because it doesn't work, and award themselves a medal for doing it.
It’s often the opposite at “hot” tech companies going through a massive hiring/growth phase (which I imagine Twitter was back when this feature was implemented). You hire one too many highly paid experts and then realize there’s no real practical work for them which couldn’t be done by a new college grad, so the end result is needless over-engineering.
They address this directly in the article. The previous horizontal crop was not an arbitrary design decision, it was put in place to improve the browsing of tweets by fitting more of them into the viewport. Their intent was to improve the user experience.
The ML-powered auto-crop was the engineering solution for how to deliver the best version of that.
They don’t say it in this article but I suspect the need to fit more tweets into the viewport was obviated by smart phone UI. The screen is tall and it is super easy to scroll quickly with the flick gesture. No need for short images anymore.
Those results are quite interesting. The bias is much smaller than I would have expected given how we've seen systems like facial recognition and face unlock behave.
Shouldn't they also check to see how frequently humans crop pictures to favor whites versus blacks, male versus female, and whether or not humans exhibit "male gaze" in their cropping decisions?
Going by the numbers they report all of the biases seemed relatively small. Slight favor for women over men and white over black and no evidence of male gaze - but single digit percentages in each case. I wouldn't be surprised if humans were more biased than machines given the results I saw.
That’s not the meaningful comparison, because Twitter has no control over what people do on their own, whereas their algorithm was introducing bias into photos with no recourse for the people posting the images.
I would argue that a 7% preference for White women over Black women is pretty darned significant. But beyond that, cropping a photo in a way that disrespects the intent of the poster is not the right approach. Fortunately, Twitter realized that it goes beyond the numbers.
How is that not a meaningful comparison? If your intent is to reduce bias in image cropping this research doesn't even suggest that changing to manual cropping would do that. What if humans are even more biased than the algorithm?
It seems like Twitter, and you, are saying they should try to minimize responsibility for bias in image cropping. Right now Twitter is responsible for it because their algorithm does it. If they foist the responsibility to users, then users will be responsible, but is there more or less bias in image cropping? No idea.
I think, if bias in image cropping is a problem, then they should check to see whether their change increases or reduces the problem. If bias in image cropping isn't a problem, then who cares. If their problem is that they don't want to be responsible for bias in image cropping then of course making the users responsible is the right way to go.
It's not meaningful because the question here is not whether the algorithm reduces overall bias. The question is whether the algorithm introduces unwanted bias into the presentation of people's posts, with no ability for them to opt out.
The bar would be different, in my mind, if we let people choose whether to use this algorithm or opt out.
Also, as Twitter finally came to at the end of their post, the much bigger issue is that this algorithm doesn't respect people's desire to curate their own public image. Taking out the bias aspect, if I post a group photo that includes my best friend, and the algorithm crops them out of the thumbnail entirely, then that's going to be really frustrating to me. If I want to center a person of color or a woman (or an LGBTQ individual, which we don't know how the algorithm would rate - we don't have the data), and the algorithm makes a different choice, then my message is damaged.
141 comments
[ 3.0 ms ] story [ 191 ms ] threadThis seems like it should have been a foregone conclusion. What was the driving force in the first place to think cropping images with an AI model was desirable? Seems like ML was a solution looking for a problem here, and I'm glad they've realised that.
Twitter crops photos to fit their preview formats. It seems like an obvious improvement to show people's faces when cropping, etc.
So, previously, the preview was just cropped in the center. But this made some images look funny, since people's faces would get chopped off.
Coming up with a workable solution to this with ML is not especially hard. You can get things like face detection off the shelf, maybe just tell your autocropper, "crop closer to the face" and have a demo within a couple days (and then much more effort to productionize it). From there, you can start introducing ML models to improve on your basic face detection. (I'm not counting face detection as ML.)
This is not a case where some massive ML model is being brought in to save two seconds of your time. This is a very natural and obvious application of ML, at a company which already does ML at scale, in a way that sounds like it has a good chance at improving the appearance of the site without introducing additional friction.
Instragram gets around this by encouraging everyone to take square photos.
The parent comments are basically saying that, just not in such an exaggerated way. "Give the users an image cropper, and be done with it."
https://blog.twitter.com/engineering/en_us/topics/infrastruc...
All those examples show large improvement. Of course they might cherrypick images with large improvement for their blog advertising the feature. But still, it illustrates why people would think it's a good idea.
Of course they don't seem to consider the idea of not cropping at all.
But "learnings" is literally, exactly, just a synonym for "lessons." Can we not?
In my experience, there's a tendency toward folksiness in certain varieties of corpspeak that causes rejection of "formal-sounding" terms and repurposing of "plainer" forms to create new words, hence lessons = learnings, protégé = mentee, and so on.
Do you think "cookings" is a word? It would be "things that have been cooked" I suppose, on that basis. I would say if it's a word it's one I've never heard before.
On the other hand, "cuttings" is clearly a word in common usage in horticulture.
It bugs me because it emerged out of tech executive culture and caught on in large part because others used it as an aspirational signal to sound like they are members of that group.
"Learnings" to me sounds like a mistake a non-native English speaker would make, perhaps by incorrectly extrapolating from the word "teachings". It's a forgivable mistake, but I don't see why we need to adopt it into the native vernacular.
(Yeah, yeah, I know, language evolves, prescriptive versus descriptive etc.. In a world where "literally" no longer means literally, I'm not going to waste much time getting angry about "learnings".)
You don’t even have to add extra required steps, generate the initial thumbnail automatically, then allow the user to move the focal point or zoom it.
Sites that allow uploading profile photos have had this type of functionality for more than a decade.
Edit: I realize I missed an opportunity for a funnier way to make this observation, which is my curse.
The blog post mentions several times how ML might not be the right choice for cropping; but their conclusion was...to keep using ML for cropping. I hope someone got a nice bonus for building the model!
If you go attach an image to a new tweet, you will see it gets uploaded right then and there, whether or not you ever hit post. Who knows what happens from there. Allowing clients to demand more crop work, at will, on top of the one that is automatic seems like it would have some implications at twitter's scale. At the very least, it would probably undo some optimizations they've made around image handling.
Cute puppy nose -> click -> porn ad.
Mitch McConnell -> click -> porn ad
My takeaway from the article was that their conclusion was to remove cropping from the product, starting incrementally on iOS. (I got cropping removed on Android as well recently). That seems like the opposite of "keep using ML for cropping"?
https://twitter.com/colinmadland/status/1307111816250748933
(Note the thread displays differently now because Twitter have changed their cropping algorithm)
Originally @colinmadland was trying to post examples of how Zoom virtual background had removed his black colleagues head, however when he posted the side-by-side images (with heads) on Twitter, twitter always cropped out his colleague and just showed him, even if he horizontally swapped the image. So, while trying to talk about an apparently racist algorithm in Zoom, he was scuppered by an apparently racist algorithim in Twitter.
It was widely covered in the press at the time https://www.theguardian.com/technology/2020/sep/21/twitter-a...
As an aside, “I think plenty of people would say X” is not a very good way to phrase an opinion. It’s ok to say you would say X and argue for it, rather than ascribe the opinion to some undefined group of other people.
I am for one happy we have unneutered the word from having to reach a bar so high, it wouldn't apply to most bigotry, but it is also unfortunate for people who have not caught on and believe calling a thing racist is a damning statement of evil intent, but it really is not anymore. Or those that insist on meaning of words remaining static forever.
[1] https://en.wikipedia.org/wiki/Disparate_impact
ok we need pictures of human faces, luckily I've got all these white people here!
on edit: it was racist in result, in that it empowers a racist system, it was not racist in intention - as in the people gathering the training data probably didn't say hey how can we empower a racist system with this?
Google exploited homeless black people to develop the Pixel 4’s facial recognition AI - https://thenextweb.com/news/google-exploited-homeless-black-...
an attempt by Google to improve its facial recognition algorithms by collecting data from people with dark skin is raising further concerns about the ethics of the data harvesting - https://www.theguardian.com/technology/2019/oct/03/google-da...
The Twitter example was not a racist algorithm. It would consistently pick one head over the other, but it had nothing to do with the skin colour. It might preference the black head for some pairs, and the white head for other pairs.
In the second example people anthropomorphised the algorithm. They assumed that any example of a preference for an images was due to a racial bias. It was easy to keep feeding it images to get to an input that confirmed this assumption.
Find the larger dimension of the image. Remove either the first or last row/column of pixels, based on which had less entropy. Keep repeating until the image was a square.
The most notable "bias" of this algorithm was the male gaze problem identified in the article. Women's breasts tended to have more entropy than their face, so the algorithm focused on that since it was optimized for entropy. To solve the problem, we added software that allowed the user to choose their thumbnail, but not a lot of users used it or even realized they could.
I assume they've since upgraded it to use more AI with actual face detection and so on, but at the time, doing face detection on every image was computational infeasible.
Clearly there is human-derived input in the system (otherwise... What's the point just crop randomly)
https://github.com/reddit-archive/reddit/blob/753b17407e9a9d...
But in short, it's a histogram of the values of the pixels.
Thanks for the insights!
Someone wrote and tested this algorithm, and either:
a) didn't test it on pictures of women, or,
b) didn't notice that it cropped breasts rather than faces, or,
c) didn't think that was a problem.
If they had noticed and cared, this wouldn't be the approach in use.
Maybe they weren't thinking about boobies a whole lot, tried like 5 random test images and shipped it?
Presumably there are a ton of failure modes for that algorithm, why get so moralistic and high-horse about just one?
Not doing so; just observing the facts.
If a proposed 'unsupervised' algorithm of this simplicity highlighted women's faces perfectly, but zoomed in on men's receding hairlines, it wouldn't have made it past the drawing board. Indeed, it's reasonable to believe that nobody would have noticed that it consistently worked for women. We certainly wouldn't know that this algorithm existed or be talking about it here.
We observe a bias in what is considered important to check before shipping.
The simplest case is just to pick the most central square, and then you could probably improve that by picking a standard square according the rule of thirds. Those are the naive algorithms - this choice of alternative algorithm is deliberate and isn't as naive or simplistic as you're claiming.
The algorithm is only considered useful because it appears to do better than that, on whatever examples that the developers tried (i.e. there was a business case for using it), and against other possible code.
Including, likely, pictures of their own selves. That's certainly what I would test it on, until it vaguely worked.
What 'awful lot of assumptions' do you think I am making? I don't imagine we are in disagreement about this.
Years ago reddit was, IIRC, not very staffed at all compared to their traffic. It's a pretty privileged take to say they should have done expensive QA entirely around your particular things that you care about.
If you don’t have the capacity to use new technologies without increasing harm, maybe you don’t have the capacity to use them.
And it was a naive image cropping algorithm, years ago, and not making use of any sophisticated 'new technologies'. The beauty of the algorithm is that it was a simple function that could have been written in 1975 and required no training, deep learning or any of that. If you want to talk self-driving cars, you've got a much more relevant measure of harm and I'm right there with you.
As it is, I'd say there's a disconnect between where you and those years-ago shoestring developers stand on Maslow's hierarchy of needs. They were being scrappy with limited resources, and you're mad at them for not having an amount of QA that would have seemed unbelievable to them under their resource constraints.
Right. I’m saying that not caring is the problem. We mostly agree on the facts. I’m objecting to not caring as an acceptable position.
We’re not talking about shoestring developers. We’re talking about platforms that millions of people including heads of state use to affect billions of people’s lives.
Jedberg was >5 years ago, maybe closer to 10 for this algorithm. They were relatively shoestring compared to now, and especially compared to current-day FAANG.
I’m not objecting to an algorithm that exposes me to more cleavage. I couldn’t care less. I’m objecting to an algorithm that exposes millions of men to a subtle but routine equation of men with faces, and women with breasts. It’s not a privilege for women to expect not to be casually sexualized because someone thought their algorithm was helpful but didn’t bother to find out it wasn’t.
Likewise the impact of racial bias. I’m not offended by seeing white faces. But it’s not a privileged position for POC to object to being basically erased from images because whoever developed the algorithm didn’t bother acknowledging more than one skin tone exists.
These things have real impact, not just on how it affects the people who are directly underrepresented. It also affects how the people who are overrepresented perceive and ultimately act toward them.
If we are aware of a negative principle that infests and invades every part of life, though we do our utmost to repel it and mitigate it and remediate it, how can we maintain the conviction that it is not the natural order of things, but is a product of solvable human faults?
I'm not denying the importance of anti-racism or of examples of bias as mentioned here; my point is that when faced with an all pervasive adversary, whether a personified devil or something more abstract, human minds tend to find relief in submitting to the seemingly inevitable via some rationalization.
Aha, perhaps that's the problem then.
edit: I googled "woman" and selected random pictures which showed the whole upper body, entropy summed over each row to the right: https://imgur.com/a/oVB57gu
(She has no work dress code she needs to conform to. Rarely does she wear a top as plain as any of your examples.)
So the solution all along was to give users the ability to crop their own photos. Why wasn't this the original way of doing things?
Instead of forcing a complicated algorithm into the Twitter experience, it seems to me that the solution all along was just to let users do what they do best-- make tweets for themselves. This incident strikes me as a major failing of AI: We are so eager to shoehorn AI/ML into our products that we lose sight of what actually makes users happy.
Someone wanted to do a feature so they could get promoted. Probably with some mumbo jumbo about how it reduces the number of clicks to create a tweet and thus increases revenue.
> In comparisons of men and women, there was an 8% difference from demographic parity in favor of women.
would have been clearer (and more correct) as "an 8 percentage-point difference from demographic parity". That 8 pp difference though is a 16% "relative" difference (58/50), or more starkly "The algorithm chose the woman almost 40% more often" (58/42 => 1.38). That said, the diagram in the post [1] is much easier for humans to parse and say "wow, that looks pretty far off!".
tl;dr: A number like 8% sounds like "no big deal", but 8 percentage points (on each side) is a big deal!
[1] https://cdn.cms-twdigitalassets.com/content/dam/blog-twitter...
It's disappointing that Twitter is trying to deflect blame for their failures onto "men".
[Edit - this is inaccurate]
They...aren’t. Leaving aside the question fo whether identifying the “male gaze” as a cobtributor to problems with the algorithm’s output would be blame passing when the entire context is taking responsibility for and addressing problems resulting from the choice of algorithm (hint: no, it wouldn’t), they initially identified that as a potential thing the saliency algorithm might* be biased for, constructed a test, and determined it was not a factor, so they are saying “male gaze” wasn’t a problem.
Kneejerk reactions to the use of the phrase without considering anything about what it is being used to say aren't helpful.
It's hard to believe that the bias was only 4% - there were a lot of people testing with images that they sourced themselves, and the preference for white people seemed much closer to 80-20.
The paper authors mention that their training data is from Wikidata (pictures of celebrities). I wonder if the types of photos in that dataset are meaningfully representative of the kinds of photos that people usually post to Twitter.
It's very easy to believe the bias was near-zero given you are citing highly motivated people on Twitter cherrypicking from thousands of examples and a little baffling you find that to be more credible than controlled systematic experimentation; note, for example, the extremely striking fact that the fuss completely missed the other bias they found which was several times larger - showing how totally useless people on social media are for testing these things and how they can conjure up "80-20" biases which don't exist.
One of the reference threads that identifies the issue happened on it by accident highlighting a surprising experience in another product (Zoom). Believe it or not, people who care about this stuff are not looking for stuff to complain about, we’re tired and overwhelmed. And I would hope that people who, upon discovering a vulnerability find and catalogue the ways it can be exploited, would be celebrated here.
I of course hope that the systems I use aren't racist against my loved ones. I am motivated to confirm whether or not they are, but I didn't go on to parlay my findings into an essay for clout. I gained nothing from doing this, except the knowledge that Twitter was suckier than I knew.
https://images2.minutemediacdn.com/image/upload/c_crop,h_135...
Imagine how much work, how much energy and effort, went into building an ML-based image cropping feature, just because an up-and-coming Designer decreed that a certain specific crop ratio was the most aesthetically pleasing to the human eye...
...so that years later, after countless hours of additional user research, they would just remove the feature because it doesn't work, and award themselves a medal for doing it.
The ML-powered auto-crop was the engineering solution for how to deliver the best version of that.
They don’t say it in this article but I suspect the need to fit more tweets into the viewport was obviated by smart phone UI. The screen is tall and it is super easy to scroll quickly with the flick gesture. No need for short images anymore.
Going by the numbers they report all of the biases seemed relatively small. Slight favor for women over men and white over black and no evidence of male gaze - but single digit percentages in each case. I wouldn't be surprised if humans were more biased than machines given the results I saw.
I would argue that a 7% preference for White women over Black women is pretty darned significant. But beyond that, cropping a photo in a way that disrespects the intent of the poster is not the right approach. Fortunately, Twitter realized that it goes beyond the numbers.
It seems like Twitter, and you, are saying they should try to minimize responsibility for bias in image cropping. Right now Twitter is responsible for it because their algorithm does it. If they foist the responsibility to users, then users will be responsible, but is there more or less bias in image cropping? No idea.
I think, if bias in image cropping is a problem, then they should check to see whether their change increases or reduces the problem. If bias in image cropping isn't a problem, then who cares. If their problem is that they don't want to be responsible for bias in image cropping then of course making the users responsible is the right way to go.
The bar would be different, in my mind, if we let people choose whether to use this algorithm or opt out.
Also, as Twitter finally came to at the end of their post, the much bigger issue is that this algorithm doesn't respect people's desire to curate their own public image. Taking out the bias aspect, if I post a group photo that includes my best friend, and the algorithm crops them out of the thumbnail entirely, then that's going to be really frustrating to me. If I want to center a person of color or a woman (or an LGBTQ individual, which we don't know how the algorithm would rate - we don't have the data), and the algorithm makes a different choice, then my message is damaged.