Infants And Ai Display Racial Bias

https://www.sciencedaily.com/releases/2017/04/170411130810.htm

https://www.theguardian.com/technology/2017/apr/13/ai-programs-exhibit-racist-and-sexist-biases-research-reveals

Both of these cases, I believe, demonstrate how the human mind and brain fundamentally work by using the analog logic of pattern recognition. A baby growing up in a white family would naturally acquire more trust in white faces merely because it matches the pattern of white faces with people it can trust more. Similarly, I believe, the AI the article describes is based on using the proximity of syntax to produce grammar. This was a heated debate in philosophy and the sciences for decades with Noam Chomsky championing the idea that grammar is innate, but recent studies have shown it is produced the hard way by crunching the numbers and even the brain organizes grammar physically by proximity to syntax. Both are interesting in that they show the drawbacks of relying entirely on analog logic which is more error prone, while digital dualistic classical logics are less error prone, but also less efficient and creative. What digital logic does provide is the necessary error correction for us to take more advantage of analog logic without accidentally killing ourselves or whatever in the process.

Anyway, that's my view which, of course, is biased.

 

These are not exclusive. No one, Chomsky included, is claiming that the processing of grammatical constructs occurs without a massive computational underpinning. That is in fact one of the central points of embracing a computational theory of mind. That being said people like Chomsky and Steven Pinker have generated arguments and empirical data that have consistently failed to be explained without invoking innate mechanisms. Chomsky's most famous argument is of course the argument from the poverty of the stimulus, demonstrating that the ability of children to produce language far outstrips the training they receive. Pinker's most famous research includes his work which reveals in some detail the nuances regarding the ability of children to apply linguistic rules. The precise ways in which young children learn and apply these rules seems to strongly support the existence of innate mechanism for language acquisition.

Clearly it can't all be innate because there are variations in language that go beyond mere differences in lexicon. That being said if someone were to provide an alternative explanation that explains everything that has been observed that would be remarkable.

 

I knew what the flaw in that infant study was going to be before I even clicked on that link

They were shown adult faces????

Does the mind of a 6 month old child process adult faces the same way it processes the faces of other 6 month olds? In the way adult researchers want to believe?


A totally useless study, waste of time



As for that second link, there is no such thing as AI yet, ridiculous calling it that, and is it sponsored by Google?

No its not, Google Translate is still a piece of shit

So the weights of a neural net lead us to find out language is biased, no shit, half the languages out there have different particles for female and male

The word flower is more heavily associated with female and pleasantness, again, no shit (sarcasm), oh totally flippin groundbreaking



Humans that end up being fluent in a dozen different languages, how do they do it? They fight for it, Practice everyday, working out what each word and phrase means in every one of those dozen languages every time they use that word or phrase in their everyday lives, thus those neurons in their brain end up getting mapped close together. The human brain is basically a back propagating neural net where the weights of input dependant on input from their daily lives, which includes the motivation to give a shit

How are you ever going to build a computer that gives a shit? This is why I'm sceptical about us ever truly getting a decent translator program let alone true Artificial intelligence

 

Providing compelling evidence for the existence of Santa Claus that can't yet be explained any other way... does not make Santa real.

All the empirical evidence on the brain indicates it works on pattern matching with the first five neural networks for pattern matching having already been mapped out. In addition to other findings supporting this conclusion, Roger Penrose's theory of quantum mechanics in the brain being responsible for memories and brain waves has already received its first two confirmations. What all that implies is that grammar is possibly intrinsic to existence itself because you are going down the proverbial rabbit hole, hence, explaining why grammar can sometimes appear to be inherited when the preponderance of the evidence supports the fact it is acquired by brute force crunching of the numbers. Hence, the reason that Chomsky's ideas are now history and merely pandering to the conservative scientists out there who refuse to acknowledge that quantum mechanics rule the universe. In fact, Donald Hoffman is a game theorist who studied all the neurological evidence and ran one computer simulation after another only to conclude that if the human mind and brain had ever remotely resembled anything like reality we would already be extinct as a species.

 

I think it should be done, the test, with multiple births, twins triplets etc, react to each other far different, than single births, wouldn't be done, because it would blow the tests out the water!
Watching my twin granddaughters, they are 3.5 now, and even have their own language, they don't react to others of the same age, and have no distinction of race, only age,...Children are not racist, till they see or hear people being so!

 

I used to babysit for three little boys who were all breast fed. Anytime a buxom women came to visit they would follow her around drooling the entire time. They are not being sexist or racist, but display a distinctive bias that merely reflects their experience.

 

Funny,. I was cleaning out my emails yesterday and ran across this. I have no idea why this was in my emails or who sent it, but it was....

https://www.youtube.com/watch?v=KykVJfLcfng

 

That's a great example of how toddlers might be biased, but certainly not racist or sexist or whatever.

The largest online study of its kind discovered that even people who think of themselves as unbiased about race actually are with very few exceptions. Having biases is just part of being human with the real question being how to become more accepting and embracing of our differences.

 

You seem to be profoundly confused. There is no relevant analogy between what I said and this statement. There is no compelling evidence for Santa Claus, most likely because there is no Santa Claus. Furthermore there is no reason to suspect that Santa Clause does exist, so there is no reason to go looking for evidence. That is to say there is no event that is begging to be explained by Santa. Finally if there were compelling evidence for the existence of Santa Claus that could not be explained in any other way then that would be a profound finding. You are missing the point that what usually leads a scientist to accept a theory is the fact that the theory explains and predicts a massive body of data that is difficult or impossible to explain or predict in other ways. That is one of the hallmarks of separating neat ideas from ideas that are probably true. This is a fundamental tenet of science.

Here are some more reasons why these cases are not analogous: I am talking about a phenomenon that clearly does occur. There is no doubt that people produce and understand language. I am also talking about this phenomenon occurring in a substrate that is known to exist (the brain) and that is underpinned by other phenomenon that are known to exist (computation). This conversation is also taking place in a larger context of facts such as hereditary variation and the strict mathematics of reproduction and biological evolution. As you can see there is not a single relevant similarity between what I said and your attempt to say that it is analogous to that hilarious statement about Santa Claus. You do realize of course a scientist never looks in a microscope and sees a little sign that says "You were right!". Nope it always comes down to demonstrating that a theory works better than others.

Do you really think that you are aware of all of the empirical evidence? I've studied psychology, language development, and computer science both in accredited institutions (degrees in Psych and Computer Science with high honors and employed full time as a programmer) and in my spare time for the last 13 years and I am certain that I am aware of much less than all of the empirical data. Also because I am aware of at least one counter example this statement is demonstrably false. Clearly not all of the evidence supports any one position.

Ah yes the fussiness of Quantum Physics does provide all of the ambiguity necessary for every single person in the world to claim that it decisively proves everything they believe about the world. Even if they believe mutually exclusive things, or rely on different interpretations to arrive at them. Or no interpretations at all because most people talking about Quantum Mechanics on the internet don't even understand ordinary physics, and have no mathematical skills whatsoever. Finally the idea that everything that has been demonstrated to be true about the processing of memories in the brain will be overturned by events at the Quantum level is a terribly confused idea. You would have to be severely lacking in basic information about the anatomy of the brain, the basics of computation, and all of the research conducted on memory to hold this belief. Now that is not to say that the Quantum level plays no role, only that it does not overturn well established facts elsewhere, which you seem to be claiming.

Seeing as though human language is essentially infinitely combinatorial it is extremely unlikely if not completely impossible to acquire it this way. You must also be unaware of the massive body of data demonstrating all of the things that neural nets simply cannot accomplish. Making a neural net capable of carrying out routine human tasks requires violating the design constraints of neural nets by including things like variables for saving states and identifying unique members of a group, or special sets of rules for incorporating assumptions about the world. This research of course comes from some of the most respected neural net researchers in the world, many of whom are dismayed at the extravagant claims people derive from their work.

The issue here is that you seem to be attached to an ideology rather than scientific research. You clearly accept an ideology that puts you add odds with what you view as the conservative scientific ideology. I hope you can realize how pointless it was of you to make this claim. Even if Chomsky were a right wing conservative scientist who eats babies and loves Adolf Hitler you would not be able to refute any of his claims by pointing that out. And the much more likely explanation is that Chomsky is a professional who spent his entire life investigating this and believes that the weight of his work points in that direction. Again just ordinary science. And you are free to try to refute the argument from the poverty of the stimulus any time you want. I tried and failed in college before I came to believe that it was true, and I never met anyone who succeeded, but if you did it I would read it and readily admit that I was wrong.

You believe that Donald Hoffman's simulations are such perfect representations of the world and the mind that you are willing to accept this claim? I guess that makes you his #1 fan. In fact I would reject this out of hand for being so obviously false. The opposite must be true. We are alive because our minds are accurate representations of the world we live in. I always look both ways before crossing a street and have never been hit by a car. So the image on my retina, and all of the processing that occurs after that point must be very accurate representations of the world. The same is true of mental arithmetic which would be useless if it were not in accord with reality.

Given that you thought my claim to be analogous to the Santa claim, that you seem to lack a basic understanding of how scientists work, that you claim to be aware of all of the evidence and that it supports your position, and that you try to argue against empirical and experimental data by accusing the authors of being conservative it is clear that I have wasted my time. The odds of you taking any of this seriously are clearly very low. Honestly I wrote all of this because I couldn't believe that misguided Santa claim you made.

The truly funny part is that if you read my original post I wasn't trying to disagree with you very strongly. I'm willing to accept that there are many facets to language development, including Quantum facets. However I am not willing to discard decades of phenomenal research and work. It is more likely that any future findings will augment rather than replace what has been done so far. (Another hilarious note: you must not be too familiar with Chomsky who actually completely disagrees with people like Pinker over the role of evolution in language. Chomsky actually believes something much closer to what you stated about grammar being an integral part of the world (though he grounds it in the computation of physics)).

 

The truly funny part is that you spent all this time writing this. The evidence for Santa Claus is that parents tell their children that's where their presents come from. Likewise, the evidence for you knowing what you are talking about is your word which, admittedly, you seem to enjoy rambling on about. I don't have time to waste with people who enjoy the sound of their own voice more than actually listening and conversing.

 

Well let's see. I claimed that there are many facets to language acquisition and the comprehension of grammar. I admitted that there may very well be a role for quantum mechanics in this. I disagreed with your conclusions based on my own 13 year investigation into this topic. I did so in a way that invited you to respond directly, by pointing out the well conducted experiments of Steven Pinker and the thought experiments of Chomsky and asking you to provide an alternative explanation to their findings. I did not accuse you of being liberal or conservative. I tried to redress your misunderstanding of Chomsky's work by pointing out that at least some of it is much closer to what you are claiming than you originally thought (making me think that you read something about Chomsky rather than reading Chomsky).

You claimed, without any argument, that all previous work has been overturned. You claimed to be aware of literally all of the empirical evidence and that it supports your claim. You accused people who do not agree with you of being conservative (a pointless accusation). It is also funny that you accuse me of basing my argument on my word when I clearly said it was based on things like the research conducted by Steven Pinker.

Lastly you are posting on hipforums so yes, you do have time to waste. I am a scientist in real life, so I do science at work and then come here to dick around. I wouldn't be surprised if you didn't actually read my posts, because you obviously didn't read the articles you posted. The one about AI in particular is very much at odds with your claims.The person interviewed for the article, Joanna Bryson, makes a number of statements clearly demonstrating the massive difference between neural nets and human brains, and that getting desired behavior out of a neural net involves tying it to a highly contrived system capable of carrying out processes that neural nets have never been able to accomplish.

Science is difficult. If you actually want to do it then you need to try harder. I've had quite a laugh but still willing to read anything you care to provide that would provide a counter to the research conducted by Steven Pinker, and the thousands of others doing similar research.

 

You both seem to obfuscate your posts on purpose so it sounds like you know what you are talking about.

This is nonsense talk:

"What all that implies is that grammar is possibly intrinsic to existence itself because you are going down the proverbial rabbit hole"

"Seeing as though human language is essentially infinitely combinatorial it is extremely unlikely if not completely impossible to acquire it this way."


Two small examples of prolonged sentences that dont actually say anything




Chomsky himself basically gave up, realized it was all about how those neurons were mapped and worked out he gets a lot more free holidays as a political speaker, you might as well nut hug Hawking or Steve Jobs


Back to that first article, Older infants follow the faces similar to those that have been giving them food and comfort for the last two years, how is that "racism"

And why the hell the assumption, that everyone on the planet seems to follow, that the parents (or any adult for that matter) are going to be the main influence in adaptive learning ????? and not other infants, hell even the family dog for that matter

 

Yeah, yeah, and I'm the Easter Bunny. Either you learn how to talk to people in a civil manner or they will simply continue to ignore you.

 

This claim is completely off base. Did you really have any difficulty determining what that sentence means? It is a very normal sentence written in accordance with the strictures of ordinary English expressing a very simple and well known concept that is almost universally described in exactly the same way by anyone who is remotely familiar with the subject matter. Maybe this topic is simply out of your league.

It is very common place and easily understood to refer to human language as being a combinatorial system that can produce an almost infinite number of novel combinations that are considered valid. This phrasing is found even in the most trivial descriptions of this field of study (popular write-ups, not even text books). Also the point that I was trying to make is equally simple and valid. It is not likely that a system can learn an infinitely combinatorial process through sheer unguided computation. Again this is all very simple introductory stuff explained in the language used by the people in the field.

 

You've made the mistake of thinking that the way I treat you is the way that I treat other people. I have intentionally treated you with disdain during this conversation. I have not been dishonest about it as I readily admitted I was just dicking around and getting a laugh out of you. I also have not hidden why I've shown disdain to you.

Early in this conversation you made the hilarious claim that you are aware of all of the evidence and that it is all in your favor. No good conversations start that way. I even gave you a chance to redact that. You want people to listen to you but you don't want to do any of the hard work to demonstrate to people that it is worthwhile to listen to you. In a world where there is more information than time you need to demonstrate that your thoughts are worthwhile, and you haven't.

So despite the fact that I'm intentionally jerking you around my challenges still stand. If you have any worthwhile information to post I will read it. In the absence of that, more dicking around.

 

If you are not already familiar check out Judith Harris. She has amassed an impressive set of evidence that, outside of extreme cases of parental neglect, that children's peer groups are one of the most profound influences on how they turn out. It is funny how everyone just thinks it is all the parents, as if children just automatically absorbed everything from their parents.

 

Children probably learn as much from cartoons, computer games and advertising as from parents and peer groups. This could be viewed as a profound problem if anyone actually cared. Welcome to oblivia.

 

And you have made the mistake of thinking I give a shit. You can rant and rave and post whatever you want and I will simply ignore anything you post from now on.

 

I didnt have trouble with sentence structure. It just didnt offer up anything. Both of you in this thread have penned long winded posts that can be condensed down to a short paragraph to convey meaning

Yes, plenty of reading over the last couple decades on the subject of early childhood development to try and figure out why I turned out so different after being raised in a mind numbingly average and normal suburban cul de sac

There are still a couple main bottlenecks that may always prevent people putting all the pieces of the puzzle together:

1. No one on the planet can remember being two years old, we are only guessing at the thought processes involved

2, We cant correlate data on self reporting of two year olds if they cant even talk yet

3. Context between scientists, a developmental psychologist will see the same problem differently to someone like Chomsky, who as you pointed out will just put more weight into his own agenda, or the agenda of whichever university at the time

4. And probably the most important one, The sum effect of ALL influences, environmental factors, weighing the importance of each, may indeed be impossible to figure out


All that in regards to the study in that first article:

There is a surge in oxytocin levels to "switch off" the babys brain during labour, and at the same time their are oxytocin recpeptors in the mothers septum so she is more recipient to their babies smell. Something that these researchers may not even know about, We dont understand how 2 yr olds see faces, they wouldnt process the face the same way an adult would and dont get the same rewards, the hippocampus isnt developed till around 2 1/2 yrs. Show pictures of adults in the same ethnic group and the infants are happier, and one can eliminate all other factors at the same time, no

But then to come out and say this proves inherant racism is ridiculous - that part is chiefly about those researchers getting attention for their own work and the university