While Humans can be seen as the most intelligent, we are hyper focused on "human way of thinking" in a way that we lose our "basic" instincts and abilities that other animal have.
My dog can understand my voice tone and emotions way better than I can understand hers, also animals can understand the difference sounds we make (words) that affect them, way better than our understanding of animals sounds.
Don't get me wrong we can make tools and we can experiment and be able to suppress all other animals. But a solo, "naked" human is like an office worker in world of manual labor.
Depends on your life experiences and working environment.
If you have worked in prisons and places with a lot of physical violence you can (some don't) acquire a distinct and accurate sense for emotion and threat, based on sound and body language.
The actual words don't matter so much, but the interaction of tone, distance, stance etc, they tell you a huge amount. People can be saying "no" and be just asking a question or pleading their case, and they can be saying "yes" and mean "i want to kill you". I used to follow the tone, and when it was going to end badly, make sure I was standing behind the person who was about to start violence (being responsible for physical security in that environment), just as it was about to kick off... Pretty grim work. But yes, you can use your intelligence to learn that stuff. Don't need to be a puppy.
Yeah, I think we spend so much of our childhoods, if they are healthy, learning to disregard those signals. Authority figure yelling but will not hurt us. Trust.
We rational humans overthink our first instinct and even learn to ignore it. And it helps us function in traditional society.
If you had a 5 meter tall dog master family responsible for giving you food, shelter and keeping you safe from dog-driven cars, you'd be quite good at reading dog expressions too.
Who had bred you to be good at reading their emotions. Dogs not only grew eyebrows that wolves don't have, they literally evolved a new facial muscle [1]. (African wild dogs have a precursor to it. But it's morphologically distinct in dogs.)
I don't think that it's as easy a two-way-street as that makes it seem.
humans developed extraordinary nuanced expression that most other animals cannot get near matching the depth and breadth of primarily because of the role that those expressions play in society itself.
an aardvark never has to put on a front so that their children aren't taken away by CPS. A donkey's livelihood is never reliant on them selling a poorly maintained used car to a sucker. Rhinos don't run for mayor.
BUT when you start considering animals that do seem to have a culture and society -- take for instance bonobos -- you start to see increased depth and breadth of expression and emotional response.
That signals to me that (most) emotions and expression come after the point of them playing a larger role than just familial maintenance. They seem to be largely reinforced by the needs of a growing social network that uses them to determine individual roles and the prioritization of 'maintenance of the group' rather than the individual.
that said, I am sure your example individual would be good at reading the expressive state of the 5 meter tall dog; I just contend that the emotional states of that dog are more simple and straight-forward than his human pet -- although maybe not since these dogs drive cars...
further research is needed on 5m car-driving dogs.
> Don't get me wrong we can make tools and we can experiment and be able to suppress all other animals. But a solo, "naked" human is like an office worker in world of manual labor.
Nah most of it is nurture. Raise a human in the wild and he'll be more in tune with nature. We have become alienated from the environment we evolved in and that's why you feel like a "naked office worker" on your own planet despite being the result of billions of years of adaptation.
Most humans simply ignore animals when they communicate. Both because they're ignorant and because they won't bother to listen. You can't expect an animal to talk with human words, but they talk all the time. Pets actively have conversations with us.
Plus there's this hardwired notion in our culture that humans are inherently superior to all animals but that's a very self-centered and short-sighted understanding of the world. We are more intelligent, yeah, but that's about it.
>We have become alienated from the environment we evolved in.
Before we start painting with all the colors of the wind too much in this thread, this is not entirely a bad thing. We are removed from stressors such as 'being eaten by large predators' and 'dying of infections from wounds'. There is a lot of 'nature' that out ancestors would be quite happy to be 'alienated' from.
Sure, but this completely neglects the aspects of humanity we left behind moving into sedentary communities. We will be forever blind to what we lost, and the morons among us will claim we lost nothing.
Which parts are you referring to? The part where you kill a guy with a rock if he looks at your wife funny?
We always lose something when evolving, that's okay. You can keep living in whatever way you want to, as long as it doesn't disrupt the liberty of another person. If you're mad that the world embraced secularity over spiritualism, or that men aren't fist-fighting for resources, blame yourself for not modernizing. Without any serious examples, your comment basically just reads like a trad dogwhistle.
When someone gives this opinion, I think to myself they surely must be terrified of a stranger killing their wife with a rock (or merely desiring it themselves). How else can you excuse or explain such an impoverished imagination?
If you are actually looking for an answer and not simply to comfort yourself, I heartily recommend Shaman by Kim Stanley Robinson.
>We always lose something when evolving, that's okay.
no, it's not necessarily okay. Nothing to do about it, and we can't change a thing , but we're not guaranteed more success in the future, and plenty of genetic mishaps occur that aren't okay -- so not every change is guaranteed to be 'okay'.
>Which parts are you referring to? The part where you kill a guy with a rock if he looks at your wife funny?
you can frame it both ways.
You say we lost the ability to bludgeon each other with rocks. (we haven't..)
I say we lost the freedom to live without social caveats like "...as long as it doesn't disrupt the liberty of another person.".
I am happy with our trajectory, but it takes a supreme blind eye to ignore that we've taken forks in the road that we can't walk back on; and i'd contend that most of what we're talking about here has nothing to do with biological evolution and everything to do with social evolution and progress.
Many of us still have the drive towards violence upon sleights, this hasn't been somehow removed by evolution -- social evolution and culture however came about and had some strong words to say about murder and what shall happen to murderers.
Social evolution used the traits of self-preservation and bent the blade backwards to provide compensation towards society for an actors bad motives. Murder didn't somehow get bred out, it became a penalized act -- those with self-interest then began to avoid penalized acts. We still have plenty of murder.
Furthermore, parent that you are replying to explicitly says that we'll be blind to what we've missed. They're absolutely right. You can oversimplify that statement into meaning cave-men headbashing each other, but there is a lot more in that statement than you seem to be willing to unpack.
We had a field where any form of philosophy, governance, or religion could have taken hold and we chose a single trajectory for our future. Of course we had to do that, that's how things work -- but the field of choices initially was so immense that it seems in poor taste to presume what parent must have meant was just barbarian violence.
Why not be altruistic in interpretation here and presume the parent wanted to speak on the difference between sedentary and nomadic life-styles?
Time and time again the issues our society faces prove to us that we have yet to find an optimal way to do things. No one can really say whether or not we're getting there faster or slower than another path would have taken us..
Rejecting the systems that have trained us into being happy to dedicate ourselves towards a series of communal goals that we have had no input in deciding, nor any means of avoiding; 'growing' the individual to identify with their governance system and the social infrastructure over oneself; are these systems an individual should want to identify with?
From a moral perspective, is it the right thing to dismantle de-humanising systems?
I agree, I am not saying that it is not nurture, in fact the opposite. Hyper focused on "human way of thinking" is not something you are born with, you adapt to it. In fact if you don't do it early you will never be able to 100%, in a way we rewire our brain to cope with how we want it to operate to be able fit in.
Our language for example, requires to be "forced on us" from early stages or you will never be able to "get it" [1]
> Most humans simply ignore animals when they communicate. Both because they're ignorant and because they won't bother to listen. You can't expect an animal to talk with human words, but they talk all the time. Pets actively have conversations with us.
With my dog I can understand angry/playful/sad/afraid/(give me food) barking/sounds and especially body language. But hearing "dog words" in random barking? Impossible.
> With my dog I can understand angry/playful/sad/afraid/(give me food) barking/sounds and especially body language. But hearing "dog words" in random barking? Impossible.
No what I meant is that animals use body language, smells and all kinds of non-auditory cues in order to communicate. And they rely a lot on behavioral reinforcement in order to communicate efficiently with us.
They don't really use verbal language like we do and therefore they also lack the tools that are required for abstract conversation.
They don't talk about complex topics like astrophysics. They usually just talk about their immediate needs but they can also convey more complex emotions like trust and guilt -- displaying a rudimentary theory of mind.
Regardless, there are many interactions you can have with your pets that entail a string of questions and answers.
Example:
1. My cat comes up to me and sits there staring. She means: "I need something of you, but I can wait"
1a. If I don't get up in a while, she will come closer and bump my leg. She means: "come on, please"
2. I ask her what's up and get up from my chair to signal that I am ready
3. She recognizes this signal, having seen it many times before, and heads for the bowl/the door/the balcony/the cat tree depending on what she needs
4. I understand her need and give her what she wants
5. She trills or purrs to tell me that I'm on the right track/my assessment is correct/to thank me
That is clearly a conversation, albeit a simple one.
Lately she's become addicted to bird and mice videos on Youtube so she comes up to me and stares intently at my laptop and/or desktop until I put those on for her.
We can understand animal emotions (mammals at least) reasonably well if we choose to try, and even in the absence of understanding, we can assume, for example, that caged animals would rather be elsewhere. But self-interest (or economic interest) often makes it more convenient to ignore what's obvious.
ASL seems like a really hard problem. I've learned a fair bit myself being friends with a few interpreters and deaf people and two different people can sign the same thing in a way that would look very different to an AI. Sometimes it feels like you have to put a fair bit of effort into understanding how a specific person signs (from my very inexperienced perspective). I'm curious if this could be overcome with sufficient data, but where is this massive archive of videos of sign? There are some databases certainly, but nothing close to the level of which we have written and spoken english via the internet. Plus then you get into regional and cultural dialects... i think banks will be obligated to hire interpreters for the forseeable future.
I'm very aware of these archives. They are very clearly not sufficient to train a model to interpret asl to the standards of corporate america. Let alone the standards of an arbitrary human. Hence my above post.
After a year-long experiment, my professor's experiment at training AI on horse races found that it could fairly accurately predict podium winners - it hyperfixated on red shirts. Happened that a rather famous jockey usually wore a red shirt. Him being in a race was a good indicator that he'd win it.
This is a thing. Going to the races and watching horses warming up in a paddock is very informative to those in the know. The first time i went, my girlfriend's father pointed out a horse that none if the others would run past, and that horse won it's heat.
I wonder if it's possible that what is actually happening is that human ability to understand animals is not optimized towards "understanding objectively what the animal is feeling" so much as it is optimized towards "understanding how to get them to do what we want."
Assuming that a significant portion of our ability to understand animal behavior comes from evolutionary instinct or ancestral folklore, then it seems reasonable that the result might be highly pragmatic. For example, our ancestors may have only cared about identifying whether a dog was communicating sufficient submissiveness to indicate that it would follow orders. Whether that submissiveness came from love or fear of punishment may not have been important.
Yeah, I came here to say this. If the AI is better at interpreting the animal's true emotion than humans, then how can humans possibly verify that the AI was closer to the actual emotion.
> “But it turns out there aren’t that many pictures of dogs and cats and sheep on the internet”—at least not ones where it’s clear how the animals are feeling, he says.
Sheep I can understand -- most people online are urban and the people who upload the most are urban -- but it also was difficult to find relevant pictures of housepets?
For some reason a lot of people seem think animals are smiling, when they're making what seems like a pretty obvious distressed face, in a situation one would imagine is distressing.
I’m not particularly happy to read about researchers intentionally causing distress to animals in order to study the response but the documentation of their research is a few grades above “obviously it’s distressed”. The point is to be able to determine whether they are feeling stress in situations that don’t have such obvious stressors.
Humans are clearly bad at analyzing other people emotions based on how much misunderstanding there are out there. Just look at how bad people are in relationships. Someone shared their experience in an event talking to a girl. I just listened, thinking, "Do they not realize that you were clearly in the wrong here?".
Misunderstanding will always be a fact of life. The diversity of humans makes it impossible to reliably interpret other people's emotions, especially when there is not a strong cultural context. We just need to learn to communicate openly and explicitly.
What creeps me out is that so many people have zero self-awareness that there is a difference between what has been communicated and their interpenetration of it.
They will be in a bad mood and conclude the text they just received must have been written in a very rude "tone".
They see your face and conclude you must be angry at them.
They take their subjective interpretation as the same as the objective truth and absolute hate to be challenged. The believe themselves to be "empaths" and "good communicators".
If you think you are very accurate at understanding other people's emotions, you are not. That is not possible. The inside state of people can not be measured by looking at their outside expression, you can only make predictions. You have to ask people how they feel.
But I think what you're talking about is someone's ability to simulate the internal state of someone else. Or its little brother: simulating it correctly and then getting mad at the result due to ego.
There’s a huge difference between analyzing emotions and being well-trained on a spectrum of behaviors. You may spend hours at a therapist to understand your own emotions, but then you go out and are expected to read people in seconds. That’s nonsense. I will even say bullshit. You may just know the social protocols better, but there’s nothing to analyze usually. It’s all common bugs in the heads of those you communicate with. The true analysis could be possible if people expressed their emotions properly, but most social games are almost designed to be as misleading as possible.
> It’s all common bugs in the heads of those you communicate with.
The problem is that people are not equipped to fix those bugs, its very hard to fix a bug with a buggy software after all. Which is why, a second opinion, in this case, an AI (or a therapist, a friend, etc) will help significantly.
I _should_ have commented on it, but I kept quite. If he really wanted to know what went wrong, I should have told him what happened, but I don't know that. If he had some ML chatbot analyzing the images and such, he would have had a second opinion.
While Humans can be seen as the most intelligent, we are hyper focused on "human way of thinking" in a way that we lose our "basic" instincts and abilities that other animal have.
My dog can understand my voice tone and emotions way better than I can understand hers, also animals can understand the difference sounds we make (words) that affect them, way better than our understanding of animals sounds.
Don't get me wrong we can make tools and we can experiment and be able to suppress all other animals. But a solo, "naked" human is like an office worker in world of manual labor.
Depends on your life experiences and working environment. If you have worked in prisons and places with a lot of physical violence you can (some don't) acquire a distinct and accurate sense for emotion and threat, based on sound and body language. The actual words don't matter so much, but the interaction of tone, distance, stance etc, they tell you a huge amount. People can be saying "no" and be just asking a question or pleading their case, and they can be saying "yes" and mean "i want to kill you". I used to follow the tone, and when it was going to end badly, make sure I was standing behind the person who was about to start violence (being responsible for physical security in that environment), just as it was about to kick off... Pretty grim work. But yes, you can use your intelligence to learn that stuff. Don't need to be a puppy.
Yeah, I think we spend so much of our childhoods, if they are healthy, learning to disregard those signals. Authority figure yelling but will not hurt us. Trust.
We rational humans overthink our first instinct and even learn to ignore it. And it helps us function in traditional society.
If you had a 5 meter tall dog master family responsible for giving you food, shelter and keeping you safe from dog-driven cars, you'd be quite good at reading dog expressions too.
Who had bred you to be good at reading their emotions. Dogs not only grew eyebrows that wolves don't have, they literally evolved a new facial muscle [1]. (African wild dogs have a precursor to it. But it's morphologically distinct in dogs.)
[1] https://www.smithsonianmag.com/smart-news/dogs-have-special-...
I don't think that it's as easy a two-way-street as that makes it seem.
humans developed extraordinary nuanced expression that most other animals cannot get near matching the depth and breadth of primarily because of the role that those expressions play in society itself.
an aardvark never has to put on a front so that their children aren't taken away by CPS. A donkey's livelihood is never reliant on them selling a poorly maintained used car to a sucker. Rhinos don't run for mayor.
BUT when you start considering animals that do seem to have a culture and society -- take for instance bonobos -- you start to see increased depth and breadth of expression and emotional response.
That signals to me that (most) emotions and expression come after the point of them playing a larger role than just familial maintenance. They seem to be largely reinforced by the needs of a growing social network that uses them to determine individual roles and the prioritization of 'maintenance of the group' rather than the individual.
that said, I am sure your example individual would be good at reading the expressive state of the 5 meter tall dog; I just contend that the emotional states of that dog are more simple and straight-forward than his human pet -- although maybe not since these dogs drive cars...
further research is needed on 5m car-driving dogs.
> Don't get me wrong we can make tools and we can experiment and be able to suppress all other animals. But a solo, "naked" human is like an office worker in world of manual labor.
Nah most of it is nurture. Raise a human in the wild and he'll be more in tune with nature. We have become alienated from the environment we evolved in and that's why you feel like a "naked office worker" on your own planet despite being the result of billions of years of adaptation.
Most humans simply ignore animals when they communicate. Both because they're ignorant and because they won't bother to listen. You can't expect an animal to talk with human words, but they talk all the time. Pets actively have conversations with us.
Plus there's this hardwired notion in our culture that humans are inherently superior to all animals but that's a very self-centered and short-sighted understanding of the world. We are more intelligent, yeah, but that's about it.
>We have become alienated from the environment we evolved in.
Before we start painting with all the colors of the wind too much in this thread, this is not entirely a bad thing. We are removed from stressors such as 'being eaten by large predators' and 'dying of infections from wounds'. There is a lot of 'nature' that out ancestors would be quite happy to be 'alienated' from.
Sure, but this completely neglects the aspects of humanity we left behind moving into sedentary communities. We will be forever blind to what we lost, and the morons among us will claim we lost nothing.
> morons
Ah yes, that old chestnut. “Anyone who disagrees with me is a dum dum”
How persuasive your arguments are.
do you have a better word to describe people who explain that they know every fork in the road that humanity has taken so far was the best choice?
Best they can do is point at our current existence -- something that individual forks may not have ever had the ability to change the outcome of.
I agree that 'moron' is a bad choice - this type of bad actor we're describing isn't as innocent as a moron.
Which parts are you referring to? The part where you kill a guy with a rock if he looks at your wife funny?
We always lose something when evolving, that's okay. You can keep living in whatever way you want to, as long as it doesn't disrupt the liberty of another person. If you're mad that the world embraced secularity over spiritualism, or that men aren't fist-fighting for resources, blame yourself for not modernizing. Without any serious examples, your comment basically just reads like a trad dogwhistle.
When someone gives this opinion, I think to myself they surely must be terrified of a stranger killing their wife with a rock (or merely desiring it themselves). How else can you excuse or explain such an impoverished imagination?
If you are actually looking for an answer and not simply to comfort yourself, I heartily recommend Shaman by Kim Stanley Robinson.
>We always lose something when evolving, that's okay.
no, it's not necessarily okay. Nothing to do about it, and we can't change a thing , but we're not guaranteed more success in the future, and plenty of genetic mishaps occur that aren't okay -- so not every change is guaranteed to be 'okay'.
>Which parts are you referring to? The part where you kill a guy with a rock if he looks at your wife funny?
you can frame it both ways.
You say we lost the ability to bludgeon each other with rocks. (we haven't..)
I say we lost the freedom to live without social caveats like "...as long as it doesn't disrupt the liberty of another person.".
I am happy with our trajectory, but it takes a supreme blind eye to ignore that we've taken forks in the road that we can't walk back on; and i'd contend that most of what we're talking about here has nothing to do with biological evolution and everything to do with social evolution and progress.
Many of us still have the drive towards violence upon sleights, this hasn't been somehow removed by evolution -- social evolution and culture however came about and had some strong words to say about murder and what shall happen to murderers.
Social evolution used the traits of self-preservation and bent the blade backwards to provide compensation towards society for an actors bad motives. Murder didn't somehow get bred out, it became a penalized act -- those with self-interest then began to avoid penalized acts. We still have plenty of murder.
Furthermore, parent that you are replying to explicitly says that we'll be blind to what we've missed. They're absolutely right. You can oversimplify that statement into meaning cave-men headbashing each other, but there is a lot more in that statement than you seem to be willing to unpack.
We had a field where any form of philosophy, governance, or religion could have taken hold and we chose a single trajectory for our future. Of course we had to do that, that's how things work -- but the field of choices initially was so immense that it seems in poor taste to presume what parent must have meant was just barbarian violence.
Why not be altruistic in interpretation here and presume the parent wanted to speak on the difference between sedentary and nomadic life-styles?
Time and time again the issues our society faces prove to us that we have yet to find an optimal way to do things. No one can really say whether or not we're getting there faster or slower than another path would have taken us..
I'd argue we've gone far beyond alienated. We're actively rejecting and dismantling, en-masse, the very systems we foundationally operate within.
Rejecting the systems that have trained us into being happy to dedicate ourselves towards a series of communal goals that we have had no input in deciding, nor any means of avoiding; 'growing' the individual to identify with their governance system and the social infrastructure over oneself; are these systems an individual should want to identify with?
From a moral perspective, is it the right thing to dismantle de-humanising systems?
-climate change
- physical health (microplastics, pollution etc)
- loneliness and mental health epidemic
- unknown unknowns of currently tech
4 off the top of my head. I'm sure I can think of more.
I agree, I am not saying that it is not nurture, in fact the opposite. Hyper focused on "human way of thinking" is not something you are born with, you adapt to it. In fact if you don't do it early you will never be able to 100%, in a way we rewire our brain to cope with how we want it to operate to be able fit in.
Our language for example, requires to be "forced on us" from early stages or you will never be able to "get it" [1]
> Most humans simply ignore animals when they communicate. Both because they're ignorant and because they won't bother to listen. You can't expect an animal to talk with human words, but they talk all the time. Pets actively have conversations with us.
With my dog I can understand angry/playful/sad/afraid/(give me food) barking/sounds and especially body language. But hearing "dog words" in random barking? Impossible.
[1]https://en.wikipedia.org/wiki/Language_deprivation_experimen...
> With my dog I can understand angry/playful/sad/afraid/(give me food) barking/sounds and especially body language. But hearing "dog words" in random barking? Impossible.
No what I meant is that animals use body language, smells and all kinds of non-auditory cues in order to communicate. And they rely a lot on behavioral reinforcement in order to communicate efficiently with us.
They don't really use verbal language like we do and therefore they also lack the tools that are required for abstract conversation.
They don't talk about complex topics like astrophysics. They usually just talk about their immediate needs but they can also convey more complex emotions like trust and guilt -- displaying a rudimentary theory of mind.
Regardless, there are many interactions you can have with your pets that entail a string of questions and answers.
Example:
1. My cat comes up to me and sits there staring. She means: "I need something of you, but I can wait"
1a. If I don't get up in a while, she will come closer and bump my leg. She means: "come on, please"
2. I ask her what's up and get up from my chair to signal that I am ready
3. She recognizes this signal, having seen it many times before, and heads for the bowl/the door/the balcony/the cat tree depending on what she needs
4. I understand her need and give her what she wants
5. She trills or purrs to tell me that I'm on the right track/my assessment is correct/to thank me
That is clearly a conversation, albeit a simple one.
Lately she's become addicted to bird and mice videos on Youtube so she comes up to me and stares intently at my laptop and/or desktop until I put those on for her.
> We are more intelligent, yeah, but that's about it.
You say that like it's not the defining characteristic of our power over the natural world
Misleading HN title. The article says AI can predict stress better than humans and only poses the question of general emotions.
Correction - AI can interpret videos of animal emotions (recorded under controlled conditions) better than humans (allegedly).
We can understand animal emotions (mammals at least) reasonably well if we choose to try, and even in the absence of understanding, we can assume, for example, that caged animals would rather be elsewhere. But self-interest (or economic interest) often makes it more convenient to ignore what's obvious.
But AI is currently failing at American Sign Language, notably at facial expressions, miserably.
Yes, we are working on that but I see writing on the wall and it is not soon enough.
ASL seems like a really hard problem. I've learned a fair bit myself being friends with a few interpreters and deaf people and two different people can sign the same thing in a way that would look very different to an AI. Sometimes it feels like you have to put a fair bit of effort into understanding how a specific person signs (from my very inexperienced perspective). I'm curious if this could be overcome with sufficient data, but where is this massive archive of videos of sign? There are some databases certainly, but nothing close to the level of which we have written and spoken english via the internet. Plus then you get into regional and cultural dialects... i think banks will be obligated to hire interpreters for the forseeable future.
Gallaudet University has the video archives.
Here is one of many:
https://media.gallaudet.edu/channel/Videolibrary+Archive/158...
I'm very aware of these archives. They are very clearly not sufficient to train a model to interpret asl to the standards of corporate america. Let alone the standards of an arbitrary human. Hence my above post.
I bet someone will turn that to Horse race betting then based on the look of the horses before/at the start of the race.
After a year-long experiment, my professor's experiment at training AI on horse races found that it could fairly accurately predict podium winners - it hyperfixated on red shirts. Happened that a rather famous jockey usually wore a red shirt. Him being in a race was a good indicator that he'd win it.
This is a thing. Going to the races and watching horses warming up in a paddock is very informative to those in the know. The first time i went, my girlfriend's father pointed out a horse that none if the others would run past, and that horse won it's heat.
I guess in theory you could maybe apply that to many things. Horse racing, poker, the courts system, commerical pilots.
Hard to imagine this hasn't already been done for decades. What role would AI play?
Maybe if we spent more time learning how to interpret animal emotions than we do building AI to do it for us the title would read vice versa.
I wonder if it's possible that what is actually happening is that human ability to understand animals is not optimized towards "understanding objectively what the animal is feeling" so much as it is optimized towards "understanding how to get them to do what we want."
Assuming that a significant portion of our ability to understand animal behavior comes from evolutionary instinct or ancestral folklore, then it seems reasonable that the result might be highly pragmatic. For example, our ancestors may have only cared about identifying whether a dog was communicating sufficient submissiveness to indicate that it would follow orders. Whether that submissiveness came from love or fear of punishment may not have been important.
If humans don’t understand animals how do we evaluate whether the language model does..?
Yeah, I came here to say this. If the AI is better at interpreting the animal's true emotion than humans, then how can humans possibly verify that the AI was closer to the actual emotion.
> “But it turns out there aren’t that many pictures of dogs and cats and sheep on the internet”—at least not ones where it’s clear how the animals are feeling, he says.
Sheep I can understand -- most people online are urban and the people who upload the most are urban -- but it also was difficult to find relevant pictures of housepets?
This type of headline is like: 'computers are better than humans at chess!' Like... isn't this something obvious by now?
For some reason a lot of people seem think animals are smiling, when they're making what seems like a pretty obvious distressed face, in a situation one would imagine is distressing.
Doesn’t take a phd to know that those pigs are unhappy.
I’m not particularly happy to read about researchers intentionally causing distress to animals in order to study the response but the documentation of their research is a few grades above “obviously it’s distressed”. The point is to be able to determine whether they are feeling stress in situations that don’t have such obvious stressors.
I wonder if they trained the systems on the times when theyre lowered?
Takes a certain kind of person to care, though.
Humans are clearly bad at analyzing other people emotions based on how much misunderstanding there are out there. Just look at how bad people are in relationships. Someone shared their experience in an event talking to a girl. I just listened, thinking, "Do they not realize that you were clearly in the wrong here?".
Misunderstanding will always be a fact of life. The diversity of humans makes it impossible to reliably interpret other people's emotions, especially when there is not a strong cultural context. We just need to learn to communicate openly and explicitly.
What creeps me out is that so many people have zero self-awareness that there is a difference between what has been communicated and their interpenetration of it.
They will be in a bad mood and conclude the text they just received must have been written in a very rude "tone".
They see your face and conclude you must be angry at them.
They take their subjective interpretation as the same as the objective truth and absolute hate to be challenged. The believe themselves to be "empaths" and "good communicators".
If you think you are very accurate at understanding other people's emotions, you are not. That is not possible. The inside state of people can not be measured by looking at their outside expression, you can only make predictions. You have to ask people how they feel.
Reading a facial expression is one thing.
But I think what you're talking about is someone's ability to simulate the internal state of someone else. Or its little brother: simulating it correctly and then getting mad at the result due to ego.
One of the problems here is that “wrong” is not universal in places that don’t have a single, universal culture.
There’s a huge difference between analyzing emotions and being well-trained on a spectrum of behaviors. You may spend hours at a therapist to understand your own emotions, but then you go out and are expected to read people in seconds. That’s nonsense. I will even say bullshit. You may just know the social protocols better, but there’s nothing to analyze usually. It’s all common bugs in the heads of those you communicate with. The true analysis could be possible if people expressed their emotions properly, but most social games are almost designed to be as misleading as possible.
> It’s all common bugs in the heads of those you communicate with.
The problem is that people are not equipped to fix those bugs, its very hard to fix a bug with a buggy software after all. Which is why, a second opinion, in this case, an AI (or a therapist, a friend, etc) will help significantly.
I _should_ have commented on it, but I kept quite. If he really wanted to know what went wrong, I should have told him what happened, but I don't know that. If he had some ML chatbot analyzing the images and such, he would have had a second opinion.
AI hype is right at its equivalent of Radium Chocolate stage of nuclear physics today
...than untrained humans, presumably, because otherwise how would you produce the model?
Mention of an app, but no references. Does anyone know of it?
These hungry piggies are the unsuspecting beta testers of Intellipig https://pure.sruc.ac.uk/en/projects/a-face-based-automated-o...
Says who? The cows?
According to the AI right?
Then AI is ready and can into U.S presidency.