Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
AI

Scientists Try Machine Learning to Understand What Animals Say (nytimes.com) 42

Posted by EditorDavid from the woofers-and-tweeters dept.
The underground-dwelling rodents known as naked mole rats make soft chirping sounds when they meet in a tunnel. So Researchers at Germany's Max Planck Institute for Brain Research "used machine-learning algorithms to analyze 36,000 soft chirps recorded in seven mole rat colonies," according to the New York Times: Not only did each mole rat have its own vocal signature, but each colony had its own distinct dialect, which was passed down, culturally, over generations. During times of social instability — as in the weeks after a colony's queen was violently deposed — these cohesive dialects fell apart. When a new queen began her reign, a new dialect appeared to take hold. "The greeting call, which I thought was going to be pretty basic, turned out to be incredibly complicated," said Dr. Barker, who is now studying the many other sounds the rodents make. "Machine-learning kind of transformed my research...."

In recent years, scientists have begun deploying this technology to decode animal communication, using machine-learning algorithms to identify when squeaking mice are stressed or why fruit bats are shouting. Even more ambitious projects are underway — to create a comprehensive catalog of crow calls, map the syntax of sperm whales and even to build technologies that allow humans to talk back. "Let's try to find a Google Translate for animals," said Diana Reiss, an expert on dolphin cognition and communication at Hunter College and co-founder of Interspecies Internet, a think tank devoted to facilitating cross-species communication....

[H]umanity is not on the verge of having a Rosetta Stone for whale songs or the ability to chew the fat with cats. But the work is already revealing that animal communication is far more complex than it sounds to the human ear, and the chatter is providing a richer view of the world beyond our own species.... [S]cientists have shown that these programs can tell apart the voices of individual animals, distinguish between sounds that animals make in different circumstances and break their vocalizations down into smaller parts, a crucial step in deciphering meaning.
This discussion has been archived. No new comments can be posted.

Scientists Try Machine Learning to Understand What Animals Say More

Scientists Try Machine Learning to Understand What Animals Say

Comments Filter:

  • Yeah! (Score:3)

    by VicVegas ( 990077 ) on Sunday September 04, 2022 @08:36PM (#62852211) Homepage

    These fruit bats won't shut the fuck up.

  • I still get shivers thinking about crawling through the dark abandoned lab with shit weapons and no armor hearing their chittering in the dark, knowing any moment they were going to jump out of the shadows and go for my throat. And worst of all was getting mole rat disease as if being simply torn apart wasn't bad enough. But getting the robot companion at the end was worth it.

  • Context & body (Score:4, Insightful)

    by markdavis ( 642305 ) on Sunday September 04, 2022 @09:01PM (#62852263)

    >"[H]umanity is not on the verge of having a Rosetta Stone for whale songs or the ability to chew the fat with cats."

    Cats make lots of different sounds. It might be a vocabulary of only 30 things, perhaps. But it is not what my cats say as much as HOW they say it, when they say it, and the complex body language that goes with it. Cats use their body, ears, paws, tail, and especially eyes to communicate a lot, without saying anything at all... and combined with what they say, can change the meanings a lot. It is something that any long-term cat owner knows. And not all cats communicate the same way. Hell, I have been around cats all my life, and sometimes I am still mystified and surprised.

    Any machine trying to learn anything about an animal (at least one as complex as a cat) that doesn't also take those things into consideration isn't going to learn but the very basics. They just aren't like humans, who have such mastery of sound/voice, that we convey TONS of information without the need for body context (facial expression, for example).

    • It is very common to anthropize cute animals especially beloved pets, have you considered that much of what you believe your cat is saying is just you making a whole lot out of very little.

      • You don't get to be king of the beasts without being able to communicate to the others. As a forced cat owner who didn't really want cats, I have learned over the years how they do communicate just as the poster described. They have so many ways to communicate because they need to be able to convey their message to other species. I am sure I have no idea what they are saying or conveying half the time, I just know when that tail starts whipping back and forth to watch out!

        • Cats (including the housecat kind) are relatively solitary creatures; it's not clear they have a need to communicate with each other, other than saying "I'm mad" when they want the other cat to go away. (As I was writing this, one of my cats used the same vocabulary to complain when the elderly dog accidently sat on him.) And kittens mew that they're hungry, and their mothers purr when the kittens are nursing. I wouldn't expect that cats have much of a vocabulary (and by not much, I mean like an order of

          • >"Cats (including the housecat kind) are relatively solitary creatures; it's not clear they have a need to communicate with each other, other than saying "I'm mad" when they want the other cat to go away. "

            Cats are, indeed, mostly solitary creatures. As such, their behavior is quite different than, say, dogs, who are more social. But that doesn't mean they don't have the ability to communicate just as (if not more) richly.

            As I said in another post: "I can also say that cats that have been loved by and

          • >Cats (including the housecat kind) are relatively solitary creatures

            Not true.

            Left to their own devices house cats will tend to congregate into colonies of dozens of individuals for mutual aid and protection - MUCH larger than an average wolf pack of 2 to 8 individuals (they're basically family units, the largest ever recorded was only 36 individuals). Within the colony they demonstrate a wide range of complex social interactions. A lot of big cats do so as well - lion prides average ~13 members, consi

        • All I'm saying is that it is far to easy to put a human-spin on the behavior of animals, especially pets. I love my cat and I like to believe that she comes to me when I'm sad because she understands my anguish and wants to comfort me but really it could be that she has figured out that taking certain actions while I'm exhibiting certain behavior is a good way to get lots of special treatment.

          • Re: (Score:2)

            by nagora ( 177841 )

            All I'm saying is that it is far to easy to put a human-spin on the behavior of animals, especially pets. I love my cat and I like to believe that she comes to me when I'm sad because she understands my anguish and wants to comfort me but really it could be that she has figured out that taking certain actions while I'm exhibiting certain behavior is a good way to get lots of special treatment.

            So, basically like humans, then?

            More seriously, the accusation of anthropomorphism has behind it a degree of Victorian Christian thinking which says that there is a clean break between humans and animals, so any interpretation of "affection" "pain" "anger" etc. is like projecting the same labels onto your washing machine.

            In real life, however, humans are animals and the challenge is to explain why they would not have emotions and thoughts, friends and enemies just as we do. Domestic cats are social animals

        • You don't get to be king of the beasts without being able to communicate to the others. As a forced cat owner who didn't really want cats, I have learned over the years how they do communicate just as the poster described. They have so many ways to communicate because they need to be able to convey their message to other species. I am sure I have no idea what they are saying or conveying half the time, I just know when that tail starts whipping back and forth to watch out!

          If its a cat what it's probably trying to convey is "fuck off" in one form or another.

      • >"It is very common to anthropize cute animals especially beloved pets"

        Anthropomorphism is more about projecting one's own emotion/feelings and higher thought on lower species or inanimate objects. We are just talking about communication. And while cats (and dogs, and most non-human animals) do have some form of emotion (as we like to think of it), that, and higher thought, are very limited. It is probably much higher in cetaceans, and non-human primates, but I have almost zero experience with those.

        • Re: (Score:2)

          by tlhIngan ( 30335 )

          There's plenty to animal communications we don't understand. Take the humble crow, or pretty much any member of the corvid species of bird. We know they communicate many things amongst themselves because we've seen the transfer for knowledge.

          This could be as simple as providing information on how to use a tool to get something, to complex thoughts like trying to describe a person's face. And that one is very interesting to scientists because no one knows how they do it.

          Basically, if you disturb a crow enoug

          • Excellent point, though I hesitate to believe crows have sufficient vocabulary to effectively describe faces. It seems far more likely that what they actually communicate is "this guy is an asshole, help me chase him off", and the others learn to recognize him on their own.

            I'm also not sure your example of them being dumb is actually anything of the sort. Buying forgiveness for past transgressions with treats and better behavior is hardly something limited to crows, nor is it dumb. Any social animal is go

      • The hypothesis that animals don't communicate has been plenty of evidence against it, and little evidence for it. At this point you're going against the science with this kind of idea. The very story we're posting on being one example.

      • It is very common to anthropize cute animals especially beloved pets, have you considered that much of what you believe your cat is saying is just you making a whole lot out of very little.

        Only a non-cat-owner could think that.

      • Re: (Score:2)

        by nagora ( 177841 )

        It is very common to anthropize cute animals especially beloved pets, have you considered that much of what you believe your cat is saying is just you making a whole lot out of very little.

        He didn't anthropomorphise anything - he didn't even make any statement about what the cat was "saying". He reported observations which any cat owner can backup about body language and other actions and noises cats make.

        Are you saying that his cat doesn't make those noises or movements?

    • > Cats make lots of different sounds. It might be a vocabulary of only 30 things, perhaps.
      > But it is not what my cats say as much as HOW they say it, when they say it, and the
      > complex body language that goes with it. Cats use their body, ears, paws, tail, and
      > especially eyes to communicate a lot, without saying anything at all... and combined
      > with what they say, can change the meanings a lot.

      Over 25 years ago, I owned an early generation Bengal, who we got when he was about 15

  • Don't bother with dogs (Score:5, Funny)

    by 93 Escort Wagon ( 326346 ) on Sunday September 04, 2022 @09:05PM (#62852269)

    Speaking as a dog owner, I can confirm that Gary Larson nailed this decades ago [pinimg.com].

  • What Animals Say? (Score:3)

    by Arzaboa ( 2804779 ) on Sunday September 04, 2022 @09:27PM (#62852317)

    Here we go. Don't need AI for this. They say a few things. Here are the translations.

    Types of animals:
    Squirrels:
    "Nut here."
    "Danger."

    Birds:
    "Worm here."
    "Danger."

    Dog:
    "Bone here."
    "Danger."

    Party:
    "Beer here."
    "Cops."

    --
    You’re All Worthless And Weak! - Neidermeyer

  • Gary Larson explained this decades ago. (Score:3)

    by Otis B. Dilroy III ( 2110816 ) on Sunday September 04, 2022 @09:39PM (#62852345)
    Good dog
    Nice kitty.

  • Anyone got a non-paywalled link? Woulda been much better for the summary.

  • Now that would be usefull for humanity. =/

  • Well, what does the fox say?

  • "animal communication is far more complex than it sounds to the human ear"

    So having discovered some time ago that animal senses (visual, auditory, olfactory) go well beyond the range of humans it has been nonetheless assumed by biologists that animal communication must take place within the human sensory range? And further, that despite living within hierarchical colonies that communications must nevertheless be simple. And why not? Every biologist knows that despite the complex biology of most living th

  • Machine learning algorithms have to be trained. Feed it input, let it know the expected result, then let it try to recognize new patterns. Here's the problem: Do we really know the expected result we should use for training? If we train the algorithm based on our assumptions, we'll get results that match our assumptions, but those might not be the correct results.

  • Machine learning can be good at finding answers that are difficult for humans. For example, machine learning can be helpful at spotting financial trends given large numbers of inputs. This task is difficult for people because of the abstract nature of financial trends.

    Understanding animal sounds is a category of things that nature is really, really good at. This is something our brains are hard-wired to be able to interpret, at least as far as that interpretation affects our personal safety. It's hard to im

  • Even in humans something like 80% of in-person communication is nonverbal.

    The big breakthroughs will probably be in video analysis, for signals that are generally too fast for humans to see.

    There was a study (can't remember who did it now) that recorded dog interactions in slow-mo, and the communication was surprisingly rich and fast.

    I'd bet most social animals have a body language repertoire that well exceeds their vocalizations.

  • If... (Score:2)

    by jd ( 1658 )

    ...they discover Naked Mole Rats love rap music and cheese, I'm going to give up and turn myself into a blue megalomaniac arch-villain.

  • Food?
    Hungry!
    Sex?
    Sex!
    Who?
    Here!
    Kill!
    Danger!

    Also, most people do not say anything else of import, why should animals be any different.

  • How does that help? Now the AI will know what the animal says, but we still won't understand what make the AI think so.

  • Something like a rat, if it has any sort of actual 'language', isn't going to have a complex language, it'll be small and very simple.
    Whales, and other marine mammals? Maybe more complex.
  • Comment removed based on user account deletion

  • Obligatory Rick and Morty Reference: https://youtu.be/fpZZQ2ov4lc?t... [youtu.be]

Slashdot Top Deals

Understanding is always the understanding of a smaller problem in relation to a bigger problem. -- P.D. Ouspensky

Close