Picture-Sorting Dogs Show Human-Like Thought

ComputerDog writes "A new study shows they can sort photographs into categories in a similar way to humans. In experiments, dogs were shown photographs of a landscape and of a dog, and were rewarded if they selected the latter using 'a paw-operated computer touch-screen'. Later they were able to correctly categorize dogs shown on an unfamiliar background landscape. '' "

Or rather ...

[Aetuneo]

Picture-sorting Humans show Dog-like thought. Who are we to claim that dogs behave like humans? Humans behaving like dogs makes just as much sense.

So...

[Schraegstrichpunkt]

... the dogs are learning provide whatever results the higher-ups want them to provide, and are rewarded or punished accordingly?

Sounds like doing science for the U.S. government.

Whats the surprise?

[Viol8]

Any animal that couldn't tell the difference between another animal and a rock or between different types of animals would soon become some carnivores dinner or fall off a cliff. Why would anyone (least of all supposedly intelligent researchers) find this ability to differentiate objects surprising? I'd imagine you'd probably have to go much further down the evolutionary tree to find an animal that couldn't do this.

Obvious.

[SatanicPuppy]

Of course they do. I'd go so far to say that most predators should show similar tendencies. We use our sight for a lot of things that the average mutt wouldn't use it's sight for, but at the most basic level, it has the exact same function for both of us. Predator/Prey identification, basic navigation, threat recognition and response.

The examples in the article are all "A dog can tell the difference between a landscape and a dog, even if the dog is on a landscape" which just shows that, like us, their eyes are drawn to the animal before the scenery. Classic response for an animal concerned with predator/prey responses. The mountains are nice, but you have to make sure of the animal first.

The main differences in visual perception would be dealing with stuff like ranging, depth perception, night-vision, day-vision, etc...All stuff to do with the actual hardware of the eye, not in the basic ability to distinguish between two similar objects.

This should be obvious from a dog's ability to tell one person from another. I've witnessed similar behavior in herbivores as well, so I'd not be surprised to find that they had the same sort of abilities, though it would be difficult to test.

Reproduce?

[TheMeuge]

Even better, any animal that can't distinguish between members of its own species and rocks, would probably have a hard time passing such stupidity on to the next generation, no?

Re:Turing Test

[Anonymous Coward]

Dog must be first trained to give expected results using training images.
By showing new and different images, their ability to understand the contents and classify them can be demonstrated.

Re:Whats the surprise?

[eonlabs]

Along the same line of thought, have you noticed that after years of believing we're superior to all animals, we still can teach a dog to respond to english, but have little to no idea what they mean when they bark a certain way? Why not see if we could build a system that lets dogs teach what they're trying to say.

Re:Whats the surprise?

[king-manic]

Along the same line of thought, have you noticed that after years of believing we're superior to all animals, we still can teach a dog to respond to english, but have little to no idea what they mean when they bark a certain way? Why not see if we could build a system that lets dogs teach what they're trying to say.

Whimper -> sad about something
door scratch -> wants out to pee or poop
Tail wag -> contentment
angry barking -> fight or flight mechanism has gone towards the former
Excited barking -> Smells owner, food, mate, friend, or stranger
Romeo's balcony soliloquies with Juliet's -> You've done too much LSD, go lie down

It all varies depending on the temperament but if you've owned a dog it's fairly obvious what they're meaning. However it's likely ham fisted autistic wolf language. Like having a child raised without anyone to teach it to speak. Parentless Child or Dog develop it's own system of communication.

Speak for yourself...

[Chmcginn]

Along the same line of thought, have you noticed that after years of believing we're superior to all animals, we still can teach a dog to respond to english, but have little to no idea what they mean when they bark a certain way?

I know the difference between my dog's "somebody I don't know is in the yard", "Mommy is home!", and her "I have to pee!" barks.

Re:Whats the surprise?

[brusk]

Partly, it's not differentiating OBJECTS, it's differentiating PICTURES of objects. That is actually a somewhat different skill.

Re:Whats the surprise?

[Antho]

Well animals rely on much more than physical static shape to distinguish things like other animals vs rocks. There are a whole array of senses to rely on and real visual cues like style of movement, speed, etc. that have to be taken into account. I think this actually interesting indicating that dogs can genuinely tell the difference between a static image of a landscape and a dog showing that based on just shape they can really tell the difference.

Re:Reproduce?

[kayditty]

Yes, but can they do it by visual recognition alone (on a computer screen, no less)?