Crows possess higher intelligence long thought a primarily human attribute • Hacker News Board

Re: Crows possess higher intelligence long thought a primarily human attribute

#381

Post by hackinthebochs » Sat, Sep 26, 2020, 6:46 PM UTC

SonOfLilit wrote: post #211

dboreham wrote: post #207

Earlier quoted context omitted.

John Searle Octopus Test

I assume you're referring to this? https://www.aclweb.org/anthology/2020.acl-main.463.pdf Gotta say in my opinion it didn't age well for a paper published in 2020 (but pre GPT-3). In practice, we have an O that is pretty good at describing how to build weapons against bears despite having absolutely no idea what "weapon" or "bear" refers to. But also, I don't seem to understand how this relates to the discussion, so…

>Gotta say in my opinion it didn't age well for a paper published in 2020 (but pre GPT-3).

Agreed. The glaring issue with their argument falls out of their own introduction. What is the "linguistic intent" of the sentence "When was Malala Yousafzai born"? Well, it falls right out of a natural language training corpus which will, with high likelihood, be followed up with her actual birthdate! So there is some non-trivial signal within the training corpus of linguistic intent. Capturing this intent is then just a computational problem. But the objective function of predicting the subsequent word has within the space of solutions the linguistic intent of the writers of the training corpus.

Re: Crows possess higher intelligence long thought a primarily human attribute

#382

Post by miltondts » Sat, Sep 26, 2020, 7:51 PM UTC

SonOfLilit wrote: post #343

miltondts wrote: post #299

Earlier quoted context omitted.

Here is a video of pigeons playing "sokoban", "talking" and playing "ping pong": https://www.youtube.com/watch?v=QKSvu3mj-14&feature=youtu.be The main problem mentioned in the video is that pigeons don't come with a repertoire that allows them to do those tasks naturally. So, the researchers have to teach them each of the behaviors independently. The pigeons then manage to put the behaviors together, in what is argua…

I'll refer to what I assume you called the sokoban example, the "banana test" from part 2, which is the most relevant to our discussion. In it, a pigeon who has been trained that sometimes pushing stuff around is rewarded with food and sometimes climbing stuff is rewarded with food manages to solve a puzzle where it must push a box to below a banana and then climb the box to reach the banana. First of all, thanks, th…

There are at least two things I think differentiate human intelligence from other animals. One is the size of the brain. Just from this difference there is probably no way for a pigeon to ever reach human level intelligence.

The other part is perhaps more interesting. For much of human history I'm not even sure that humans could solve sokoban puzzles. In fact if you raise a human without social contact I suspect he would not be able to do it. This appears to indicate that we now have something in our environment (quite possibly what we call language) that allows us to be much more intelligent with the same DNA. Can other species with the same kind of brain complexity/power develop this? That's a very interesting question. This story [1] appears to indicate that at least raising other species as humans doesn't appear to work, but to me, this doesn't mean that there is no way of getting them to our level of intelligence, just that they don't have the same communication tools we do.

[1] - https://www.smithsonianmag.com/smart-news/guy-simultaneously...

Re: Crows possess higher intelligence long thought a primarily human attribute

#383

Post by jancsika » Sat, Sep 26, 2020, 8:09 PM UTC

SonOfLilit wrote: post #203

jancsika wrote: post #194

Earlier quoted context omitted.

> I know at least octopi can use computer screens, and I know they enjoy hard puzzles that reward them with food, so it should be easy to teach them some computer puzzle game? I'm not sure it's that easy. For example, suppose a race of hyperintelligent octopuses beam down and decide to study your intelligence. They notice in one test that if you begin scrambling eggs for breakfast and they cut off both your arms at t…

I mean, they'd be right that my intelligence is not as well distributed as theirs. But that's a specific feature of their intelligence, that's not general intelligence. If an octopus can beat me at Sokoban, I don't care whether it thinks with its brain or its arms, I would like to hire it as a programmer for my startup. If an octopus tries its best to play Sokoban, beats the tutorial but can't figure out the levels w…

> If an octopus can beat me at Sokoban, I don't care whether it thinks with its brain or its arms, I would like to hire it as a programmer for my startup.

You'd need to know something-- indeed, a lot-- about the entity in question in order to design tests to measure its general intelligence in the first place. Otherwise you're going to default to anthropomorphize the entity in question and misunderstand what it is you are measuring. E.g., vastly overestimating the language skills of a gorilla, or perhaps believing that ML algorithms somehow provide a higher level of objectivity than the data they were trained on:

https://www.wired.com/story/algorithms-shouldve-made-courts-...

Re: Crows possess higher intelligence long thought a primarily human attribute

#384

Post by Twirrim » Sat, Sep 26, 2020, 8:19 PM UTC

ende wrote: post #231

Twirrim wrote: post #166

Earlier quoted context omitted.

Years ago, we had a dead crow in our yard. We think our dog did it, but we didn't see. My wife and I dealt with it, but had a lot of crow onlookers. They were pissed. For the next year or so, every time either my wife or I would walk out the door, if crows were in the vicinity they would sound an alarm until they were happy we weren't going to attack them. They never seemed to do that for the dog, though.

That’s called “being framed” ;)

Our dog is definitely smart enough to try and pull a stunt like that.

Re: Crows possess higher intelligence long thought a primarily human attribute

#385

Post by YeGoblynQueenne » Sun, Sep 27, 2020, 8:15 AM UTC

AnIdiotOnTheNet wrote: post #377

YeGoblynQueenne wrote: post #360

Earlier quoted context omitted.

I don't think it's monstrous to eat an intelligant animal. Pigs are very intelligent and sensitive animals. And yet, I eat pigs without remorse because being born on a planet where animals eat each other is not something I chose, therefore I do not feel guilty about it. Also, it's still mosquito season in Greece where I'm currently located and I've been feasted upon often enough myself to know that refusing to feed o…

> And yet, I eat pigs without remorse because being born on a planet where animals eat each other is not something I chose, therefore I do not feel guilty about it. You were born on a planet where rape and murder[0] is common in the animal kingdom too, yet quite rightly we humans have collectively decided that they are monsterous practices. "It's the natural order" is not a good basis for reasoning. [0] which we will…

Thanks for the reply. I think you understand my comment as making a point about the "naturalness" of eating animals (I certainly did not use the expression "natural order"). I'm rather making a point about universality. As far as I can tell, all animals eat other living things- be it plants or other animals. Humans eat both plants and other animals. This is not my choice, so I don't feel morally responsible for it. It's how I found things when I came into this world.

I would rather things were different. I would rather not have to kill anything in order to feed. However, this is what we do. We kill other animals and eat them. We kill plants and eat them. Well, at least we usually eat them (https://en.wikipedia.org/wiki/Surplus_killing). Note that I find killing for sport morally unjustifiable.

Rape and murder are not universals. That is, some animals rape and murder their own and some cannibalise their own, but not all animals do. In any case, I personally have never felt a compelling need to rape or murder anyone, so I can attest to the fact that it's not a universal experience for humans to rape and murder (assuming I'm actually human and not some weird mutation eh).

Edit: As a summary of my views, I recommend "that's how it works" over "it's the natural order", which has connotations of a moral imperative that I don't recognise.

Re: Crows possess higher intelligence long thought a primarily human attribute

#386

Post by SonOfLilit » Sun, Sep 27, 2020, 10:44 AM UTC

miltondts wrote: post #382

SonOfLilit wrote: post #343

Earlier quoted context omitted.

I'll refer to what I assume you called the sokoban example, the "banana test" from part 2, which is the most relevant to our discussion. In it, a pigeon who has been trained that sometimes pushing stuff around is rewarded with food and sometimes climbing stuff is rewarded with food manages to solve a puzzle where it must push a box to below a banana and then climb the box to reach the banana. First of all, thanks, th…

There are at least two things I think differentiate human intelligence from other animals. One is the size of the brain. Just from this difference there is probably no way for a pigeon to ever reach human level intelligence. The other part is perhaps more interesting. For much of human history I'm not even sure that humans could solve sokoban puzzles. In fact if you raise a human without social contact I suspect he w…

I'm not sure how much I buy the "size of the brain" limitation, given the brain-size ratio between parrots and chimps hardly reflects their intelligence ratio as far as we can tell.

I tend to believe any creature who has the capacity to learn language also has ehe capacity to solve Sokoban and vice-versa, so testing the easier to test of the two should be the first step and figuring how to teach them a complex language should be the second.

(By the way, corvids pass such complex information that some people suspect they already have language, although I would bet against. We know of cases where people the crows hated left a place, returned years later to visit and all the local crows still hated them, and it was implied where I read it that it was long enough that it was all a new generation of crows.

Re: Crows possess higher intelligence long thought a primarily human attribute

#387

Post by curation » Sun, Sep 27, 2020, 2:00 PM UTC

Intelligence is not contained wholly within an individual of a species but includes relationships with fellow species and the biosphere.

Re: Crows possess higher intelligence long thought a primarily human attribute

#388

Post by unkulunkulu » Sun, Sep 27, 2020, 5:07 PM UTC

powerapple wrote: post #300

liveoneggs wrote: post #190

Earlier quoted context omitted.

crows and other birds are very good at recognizing faces so they absolutely know it's you

Can they recognize person? I highly doubt it because even humans, we are very good recognizing faces, still not very good recognizing faces from another race. I know birds can recognize themselves, from feather patterns and so on. I would think the way they see us would be similar to the way we see them, they probably look the same to us.

Why do you think people are bad at recognizing people of another race? When you rarely meet people of some race then maybe, yes, but I think it must be a matter of experience and you will get better. Think twins for example, after a little time with them you’ll differentiate them easily.

Re: Crows possess higher intelligence long thought a primarily human attribute

#389

Post by SonOfLilit » Thu, Oct 01, 2020, 4:24 PM UTC

jancsika wrote: post #383

SonOfLilit wrote: post #203

Earlier quoted context omitted.

I mean, they'd be right that my intelligence is not as well distributed as theirs. But that's a specific feature of their intelligence, that's not general intelligence. If an octopus can beat me at Sokoban, I don't care whether it thinks with its brain or its arms, I would like to hire it as a programmer for my startup. If an octopus tries its best to play Sokoban, beats the tutorial but can't figure out the levels w…

> If an octopus can beat me at Sokoban, I don't care whether it thinks with its brain or its arms, I would like to hire it as a programmer for my startup. You'd need to know something -- indeed, a lot-- about the entity in question in order to design tests to measure its general intelligence in the first place. Otherwise you're going to default to anthropomorphize the entity in question and misunderstand what it is y…

Sokoban puzzles are very close to math-complete (in the same sense as in Turing-complete or NP-complete).

Ockham's Razor leads me to believe it's much more probable to find an alien (or animal or ML) entity good at math than an alien entity overfitted for complex Sokoban puzzles.

Now, I may undermeasure (maybe they're very intelligent but suck at spacial reasoning) but it would be hard to overmeasure (if they can do this kind of math, I'd like to hire them, period).

Re: Crows possess higher intelligence long thought a primarily human attribute

#390

Post by SonOfLilit » Thu, Oct 01, 2020, 4:25 PM UTC

filoeleven wrote: post #271

SonOfLilit wrote: post #18

Hopefully someone in the field will read this and be able to answer my question: I've spent a lot of time googling research about animal intelligence, and I now know a bunch of things intelligent animals can do, but I have no idea where their limits are, what they can't do. Everything published uses such... nonflexible tasks. If we're trying to measure general intelligence, why isn't anybody trying to teach corvids,…

Definitely a neat idea! Your local cooperative extension service should be able to get you in touch with an animal behavior expert, since they are often affiliated with colleges. Thinking about the project with corvids in mind, I’d suggest two nuances to your protocol. 1. Stick with plain mazes that have no crates in order to train the association with the onscreen character. Basically just spending more time with na…

Thanks, these are great suggestions, I'll make sure to use them!