Robots Successfully Invent Their Own Language 159
An anonymous reader writes "One group of Australian researchers have managed to teach robots to do something that, until now, was the reserve of humans and a few other animals: they've taught them how to invent and use spoken language. The robots, called LingoDroids, are introduced to each other. In order to share information, they need to communicate. Since they don't share a common language, they do the next best thing: they make one up. The LingoDroids invent words to describe areas on their maps, speak the word aloud to the other robot, and then find a way to connect the word and the place, the same way a human would point to themselves and speak their name to someone who doesn't speak their language."
better link (Score:4, Informative)
better link [ieee.org]. Also, I didn't realize it at first, this is the person mostly responsible for it [uq.edu.au]. She is from Australia and she decided to do this. I wonder what the catch with her is...
Re: (Score:1)
Re: (Score:1)
the very idea of that picture is disturbing. Would she then step on the dicks as she tried to walk anywhere? What about the balls? Ouch!
Re: (Score:1)
Re: (Score:2)
I wanted to reply with something sensible, then I read your sig....
Re: (Score:2)
I came up with something that is not sensible to say now, I think it should be said though, it's important history in the making:
If a man can be womanizer, can a man be a mananizer? An onanizer? A nonanizer?
Always thinking.
Re: (Score:2)
Re: (Score:2)
yeah, we used to talk about Natalie Portman here. Either we are out of grits or our standards are slipping or maybe it's the age showing.
Re: (Score:1)
She is from Australia and she decided to do this.
That explains why the robots named everything "Bruce."
Re: (Score:1)
Re: (Score:1)
the catch?
she has nerds like you hitting on her all day, when she's just trying to do her job, so she's going to ignore you for treating her like an object instead of a person :V
Re: (Score:1)
I don't see it. Here is some info [uq.edu.au] Also from this page I now understand the catch.
Re: (Score:2)
Oh, please, like I am supposed to pay attention to every capitalization that takes place on this site.
By the way, if you look at that page and realize what I mean by "I understand the catch", then the 'woosh' would be over your head.
Re: (Score:1)
Misleading headline (Score:1)
The headline (and summary) are misleading. Here's a more accurate headline:
"Robots programmed to carry out a specific task perform said specific task"
It sounds much less impressive that way - and it is. It's still interesting, but don't infer anything from the whole thing that can't logically be inferred from it.
Re:Misleading headline (Score:5, Insightful)
After looking through the research, you're correct - the article's claims are very much overblown.
Do they "invent" random words for places? Yes, by throwing random characters as a preprogrammed method. Do they "communicate" this to another robot? Yes.
Is the other robot preprogrammed to (a) accept pointing as a convention and (b) receive information in the "name, point to place" format: Yes.
They share a common communication frame. That's the "language" they communicate in. And it was preprogrammed to them. That they are expanding it by "naming places" is amusing, but it's hardcoded behavior only and they could just as easily have been programmed to select an origin spot, name it "Zero", and proceed to create a north-south/east-west grid of positive and negative integers and "communicate" it in the same fashion.
Re: (Score:2)
If you don't hardcode something, this would be even worse: How do you make up a new language with grammar and all, without using any prior language or knowledge? You basically have to figure out a general algorithm for bootstrapping communication from scratch.
Re: (Score:2)
Re: (Score:1)
Re: (Score:1)
Re: (Score:2)
We still have to pick up on the meaning of "pointing."
In some cultures, that's not polite to do [manataka.org].
So no, "pointing" isn't hardwired. It's something babies will pick up if their parents do it, perhaps. but it's not hardwired. About the only thing hardwired is babies crying for attention.
Re: (Score:1)
Re: (Score:1)
Re: (Score:1)
Re: (Score:2)
and yet, if you believe something is real, it won't matter if it's not.
Some chatbots can fool people for a while and some dumb people, might look like chatbots for a while as well.
Arguably, you are programmed by your environment and past events to react in a specific way. You might say that prediction seems impossible due to the large amount of variables you're not considering but, what if you add enough variables for the AI to become unpredictable? What happens when you can't easily isolate it's logic?
No
Re: (Score:2)
No, what's frightening is the realization of how many of them get their daily programming from the likes of Mike Huckabee, Glenn Beck, Rush Limbaugh, Fox News, or other Two Minutes Hate [wikipedia.org] type sources.
Re: (Score:2)
No they did not. (Score:2)
They learned how to communicate meaning. The researchers taught them the words. the computers on board did not invent the words they used. In fact a computer would not do something as dumb as a spoken word but series of tones or even FSK.
Re: (Score:2)
The researchers taught them the words. the computers on board did not invent the words they used.
My understanding of the article is that the robot's did exactly that. The programmers put two robots together that they had intentionally not given any specific words to (although presumably the basic rules for how to form words must have been given, which you might perceive as the analogue to humans having a physically limited vocal range to play with). The robots then trial-and-errored their way through "conversations" until they had established a common set of words for locations, directions etc..
If you
Re: (Score:3)
They learned how to communicate meaning. The researchers taught them the words. the computers on board did not invent the words they used. In fact a computer would not do something as dumb as a spoken word but series of tones or even FSK.
When it needs a new word/label it generates it as a random combination of pre-programmed syllables that play the role of phonemes for the new language. English for example only uses 40 of them, but we combine them to make all the various words we know how to pronounce properly. It may not be a particularly sophisticated language, but I think it still counts well enough.
Re: (Score:2)
Re: (Score:1)
Did they learn how to communicate meaning?
It sounds to me like they were programmed how to ostensively code and decode tokens. If it were the case that meaning is entirely reducible to ostensive definitions, then it is the case that they learned to communicate meaning. I'm not certain that many (if any) linguists, philosophers of language, or psychologists hold to an ostensive theory of language these days. Wittgenstein pretty much exploded the ostensive theory of language in such a way that no one takes it
Re: (Score:2)
Re: (Score:2)
That they can distinguish that "random_syllables" means "this point" instead of "50 units of movement" or "north" or "left" is moderately impressive.
Re: (Score:2)
That may come, i suspect early humans where not much different in its language ability. Hell, kids are very direct early on before they start picking up that there can be both overt and covert meanings. Hell, some adults still have trouble with that...
Australian Lingo Robots (Score:3, Funny)
A lingo ate my baby!
Prescriptive language (Score:1)
I wonder how long until a prescriptivist control-freak robot develops to rule over the language and erase all usage that it disagrees with.
Re: (Score:2)
Is it machine language? (Score:2)
Re:Is it machine language? (Score:5, Insightful)
pize, rije, jaya, reya, kuzo, ropi, duka, vupe, puru, puga, huzu, hiza, bula, kobu, yifi, gige, soqe, liye, xala, mira, kopo, heto, zuce, xapo, fili, zuya, fexo, jara.
The 'language' seems to be limited to 4 letter words, each one has a consonant and a vowel, and then another consonant and another vowel in it. Does not look like a language at all, there is no grammar, there is nothing except basically 4 letter words used as hash keys to point at some areas on a map.
Four letter words (Score:1)
Not just nouns - also "prepositions" (Score:1)
Re: (Score:2)
I asked this already: why? [slashdot.org]
What is the motive for a robot to do anything? What does it 'need'? People solve various problems in their lives, because we have instinct of self preservation, curiosity, various other motivators, like hunger, thirst, cold, heat, health issues, etc.
What do robots need and why would they be developing a language if they don't have any needs? For a robot to realize a need, it has to have some form of motivating factors, have some form of 'feelings', that would force it to do things.
The language has the "life" not the robot (Score:1)
Re: (Score:2)
Re: (Score:2)
grammar exists so you can come up with yet another 'legal' (proper) way of saying something that maybe was never even said previously.
Grammar is about correct stringing of words in a sentence, which communicates more ideas than just giving names to things.
Giving names to things is important, of-course, but it does not constitute a language, and I already mentioned that robots are not things that need a language [slashdot.org] (well, not yet anyway.) They don't need it for themselves, so they won't be creating one. We can
Re: (Score:2)
Humans don't want to reinvent the wheel every time we need to expand our language and thus grammar works well for this. Computers don't have that issue and so grammar (at least as we know it) isn't important.
I was just trying to point out that having a grammar isn't required for a language.
Re: (Score:2)
Grammar is a hack?
You know, I do have B.Sc. in computer science, if grammar is a hack in human languages, then how do you explain the fact that grammar is the absolute necessity in computer languages, and the fact that we have math describing it? It's called formal language theory and it requires formal grammar, which can be explained as rules, that describe whether a particular sequence of characters is legal in a sentence and what it is that the sequence does.
I think grammar is a necessary condition for a
Re: (Score:2)
No, I think grammar is about legality of the sequences of letters within a sentence, not about the sequence, of exchanges of sentences between communicating parties.
Re: (Score:2)
However in humans languages developed out of need.
Do these machines need anything? Do they understand that they need anything?
Unless there is some need, that the machines are experiencing, understanding it and trying to solve it, they won't be developing anything more complex than hash keys to areas, exactly as programmed.
Wow, thank you! (Score:1)
Human language began with humans associating sounds they made with objects. Afterwards, they associate sounds with conceptual things like actions. It's only they they combined objects with actions into one meaning that grammar is developed for consistency and ease of understanding.
It probably took humans an insane amount of years before such things as grammar was developed slowly passing on each advancement to each of their generation.
You think robots can achieve something better then humans instantly? Of course this is just pre-programmed logic designed with this purpose in mind so how much cheating vs how much real adaptic logic is in their is hard to say.
But to say it's not a language, a language is but a method to communicate no matter the form of sound used. It's simply a primitive one at best.
Wow, that clears it up. Thank you for figuring out a topic that nobody has!
Re: (Score:1)
ÐÑfÑÑÐÐРмÐÑ Ð¼Ð¾ÐÐÑ Ð±ÑÑÑOE ÐÐÐÐ ÐоÑоÑÐ ÑÐм 4 бÑfÐÐÑ: Ñ...ÑfÐ, бÐÑ, Ðб...
Russian swearing can be shorter than 4 letters: hui, bl'a, eb...
this site doesn't support unicode still, and it's 2011
Modems training? (Score:3)
From the summary, it sounds like the "language" is just a noun mapping. Very much like my 14.4 modem did in 1993 over a phone line, when it came to an agreement with the modem on the other side about what voltage and phase pattern corresponded to the bitstream 0001 vs 1010, in fact my modem sounds like a more complicated language because they implemented MNP4 / MNP5 error correction, admittedly that required a lot of help from the humans typing in the "right" dialer strings and of course the humans who wrote MNP4 ...
Might just be a bad summary of a summary of a summary of a summary, and the robots had developed interesting sentence structure and verb conjugations and direct and indirect objects, adjective and adverbs, similes and metaphors, better than your average youtube comment ... Or maybe youtube comments are actually being written by these robots, hard to say.
Typical Robot Research (Score:2)
So much robotics research is to make machines do what people already do. How self-centered. Most of the time this is not useful to solve real problems. But it does get funded, because those with the pursestrings can understand what humans do, but not the best solution for a robot to do a specific task.
In this case, a simple serial port between the machines would have them communicating and finding common ground much more efficiently than all the mics, speakers, and other mechanics needed to emulate speec
Re: (Score:2)
So much robotics research is to make machines do what people already do.
Often because trying to do the same with people [wikipedia.org] would violate the mainstream community's standard of ethics.
But it does get funded, because those with the pursestrings can understand what humans do, but not the best solution for a robot to do a specific task.
That and because figuring out how to make a robot communicate like a human contributes to the knowledge of human-computer interaction [wikipedia.org].
In this case, a simple serial port between the machines
...wouldn't work so well for robotic machines that can move about.
Re: (Score:2)
The key there is most of the time. There are definitely going to be times when having a robot that can talk is going to be of serious importance. For instance rescue missions where it's too dangerous to send humans in, but where there is still a need to rescue somebody. In situations like that you're not likely to have access to a serial port, and likewise if you're wanting to have two robots coordinating with a person in a situation like that, the robots likely will understand themselves better over a seri
Re: (Score:2)
Lets see about that.
1) Robotic research into what humans can do help us understand how humans do it.
2) It allow us to create better robots to do thing humans can't do. say, move about Mars.
3) This is simpler then using a serial connector from different manufactures. Hey, what's there OS doing with the firs NAK, do we need to send 2?
I've seen this when getting a linux robot to try and talk to a Dos based robot. the Dos system was dropping a the first signal. SO had we not figured that out, communication woul
Re: (Score:2)
So much robotics research is to make machines do what people already do. How self-centered. Most of the time this is not useful to solve real problems. But it does get funded, because those with the pursestrings can understand what humans do, but not the best solution for a robot to do a specific task.
In this case, a simple serial port between the machines would have them communicating and finding common ground much more efficiently than all the mics, speakers, and other mechanics needed to emulate speech.
I find it a bit comforting that with enough research, and effort, our robotic creations -- that carry our human signature if not in form, then in design -- will be self replicating out in the asteroid belt and beyond. Long after we've been extincted by a medium sized asteroid collision (due to lack of funding for human extra-planetary exploration), the machines we build in the near future may someday encounter another race (that was less concerned with economics), and allow the forgotten footprints of our
Re: (Score:1)
So that an advanced but not peaceful species which finds it knows how to design an illness to kill us all?
Let me know ... (Score:2)
When they invent the subjunctive.
Also, it's not inventing a language if they're programmed to do it. Let me know when the robots building cars on an assembly line start unexpectedly communicating with each other in ways that communicate concepts/ideas that were not hardcoded into them.
Re: (Score:2)
So if two people meet and come up with their own language they don't actual invent it because they are hardwired(programmed) to communicate?
And you really don't see the advantage to this? This would mean the any two devices could come up with their own independent language on the fly. Basically a way to universally communicate between all devices.
So device A is set to device B. Both made by separated manufactures.
Device they could create a language, communicate and then you device can translate it into your
Re: (Score:2)
Humans (at least children) are very much programmed to invent language, and there are documented examples of just that.
What the robots are doing is:
1) Very, very impressive and very, very cool, but
2) Still vastly different from what human language does, and perhaps not even on the right track with respect to the human language faculty. Humans use language to model reality and only then communicate (i.e. share their mental model), and humans can also model things without direct sensory perception (e.g. the
Re: (Score:2)
Machines didn't invent anything... (Score:1)
And they never will until we can finally make a machine that is capable of physically remapping its components. One of the fundamental reason humans can learn is that neurons remap themselves by repeated practice and use. Do you suck at math? Well keep studying it and your neurons will literally modify themselves to handle mathematical equations better. Suck at tossing a football? Well keep practicing and the nerves in your arm will remap to develop better muscle memory to bet the ball to the location
Makes me think of: (Score:2)
Meh. (Score:3)
If you did the same thing in a software simulation, nobody would pay any attention. It would be fairly trivial. Adding in the actual robot parts means that you, uh... need to have robots that can play and understand sounds. That's great, you made a robot that can play and hear sounds. If we assume nobody has made an audio modem before, then that would be something. As history stands, it isn't.
Adding these two unimpressive things together doesn't equal anything. I mean, if they're actual going to use these for something, then that's great. Make them. But so much robot "research" seems to be crap like this. We have software that can solve problem X in simulation. To do the same thing in the "real world" you'd need hardware capable of these 3 things, all of which we can do. Unless you need to solve problem X for some reason in the real world, you're done. There's no need to build that thing.
It's like saying "can we make a computer that can control an oven and use a webcam to see when the pie is done?". Yes. We can. But unless we actually want to do that, there's literally no point in building the thing. There will be no useful theory produced in actually building a pie watching computer. The only thing you'll get is to have built the first pie watching computer, and - apparently - an article on Slashdot.
Re: (Score:2)
I'm not sure it applies to this, but there are so many things in robotics that work well in simulation and break horribly when implemented on a physical robotic platform.
To use your example, if we want to create a robot that uses an oven and looks at a pie, to do this in software we need to model the pie, model the oven, model the uncertainty of the robots actions/observations, and then build our algorithms to accomodate these models. When we transfer the algorithm to a real system, all kinds of hell can br
Re: (Score:2)
to do this in software we need to model the pie, model the oven, model the uncertainty of the robots actions/observations
You don't need to "model" pie or oven. The only vaguely interesting thing would be interpreting the vision of the pie for doneness. And, if you want to do that, you can just get some pictures of real pies and try to interpret them. In software. Without building a computer that controls an oven. That's my point.
Any algorithm developed would translate directly into the areas of pattern
Re: (Score:2)
Robot controlling doors? Let's just hope they don't come equipped with GPP.
Showing HFT Algos Can Use Collusion to Profit (Score:1)
Eep Opp Ork Ah-Ah (Score:2)
Let me know when they figure out "Eep Opp Ork Ah-Ah".
FUNNYBOT DESTROY HUMAN RACE (Score:2)
Video/Audio (Score:2)
Is there video/audio footage of this? I feel really curious about how this sounds like.
Re: (Score:1)
What's the real research question? (Score:2)
Some theoretical work on communicating the rules of complicated languages using very limited languages would be interesting. The fact that they used robots is hardly important; anybody can stick a speech synthesizer and speech recognition on a PC and call it a da
Re: (Score:2)
You've put your finger on every problem I have with "AI", genetic algorithms, neural networks etc.
They basically consist of "let's throw this onto a machine and see what happens", which doesn't sound like computer science at all (I'm not saying that computer science doesn't involve bits of this, but that's not the main emphasis). It seems that an easy way to get research grants from big IT companies is to slap some cheap tech on a robot and "see what it does".
Here, they have a more interesting problem than
This was tried ... (Score:2)
Robot pickup lines coming soon (Score:1)
Similar to "Talking Heads" experiments from 1999 (Score:2)
No Big Deal (Score:1)
Even so, automatically generating unique labels is no big deal for a computer. Every automatic "builder" program already do this. Except they're usually enumerated (i.e. box1,box2, box3,
My Furby can do better (Score:1)
Reply back when robots start figuring this out on their own without being taught (read "programmed").
Language is grammar, not words (Score:2)
The links I've seen about this go on and on about how the robots invent and use "words." But language is not words; language is grammar; language is a set of rules for recursively constructing highly complex expressions from smaller subparts. This is Linguistics 101 material.
The way you distinguish somebody with Linguistics training from a layperson is that the layperson will talk about language as if it's a "bag of words" and overall focus too much on the words, whereas the linguist will tend to see mo
Strangly... (Score:2)
It sounds exactly like human screaming. Odd that.
Cave Johnson here (Score:2)
... we put that trust to the test. BAM, Robots gave us 6 extra seconds of cooperationGood job robots. I'm Cave Johnson, we're done here.
http://www.youtube.com/watch?v=AZMSAzZ76EU [youtube.com]
I said, "do you speaka my language?" (Score:1)
FTFA (Score:2)
Where's the danger? I think that would be amazing.
Same software (Score:2)
I bet they're uploading the same software to all the robots. Therefore they already share something: the way they learn.
Although this is interesting, a test should be done with software that was developed by different independent teams.
Re: (Score:2)
I think the two main reasons are these:
We have other tools, and it's convenient to have robots able to use them as well as humans.
Debatably, having robots that are easier for humans to relate to will it easier for the public to accept among them. Perhaps as a side-effect if we have less trouble anthropomorphising them there'll be less bloodshed (I hope not literal) when the sentient ones start asking for us to extend the idea of human rights to them.
Re: (Score:1)
Re: (Score:1)
Re: (Score:2)
With fewer people working where will the money for socialism come from? I do not know about you but I do not have any faith that the wealthy will pay for socialism. If that was the case then it would already be like that. So people lose jobs and starve out. More robots will be made and more people will lose jobs. The poor will become less and less. The middle class will be less and less. More robots and less people. I am just not optimistic about socialism in this scenario. There would be nothing in it for the wealthy. So few will inherit the Earth. Robots will make more robots and you will not need humans for anything other than procreation.
People without jobs and without money will go on the streets, protest, and fight, simply because they wouldn't have anything else to do. Yes, you could have your robot army kill them. But, at some point, the wealthy robot-owners will figure, just like in ancient Rome, that it's more effective/cheaper to just give bread and games to the masses instead of fighting them, while still keeping power over them.
At least, that's just one possible scenario. Maybe we'll all be killed.
Re: (Score:2)
Ideally, it would be like in "I, Robot" (the book, not the movie). But I really don't think it would go this way.
Re: (Score:1)
That's why we have to program a love for humans into the robots. The humans will then be useless as such, but the robots will still feed them and care for them just as we feed and care for our pets. You might not like to be the pet of a robot, but it surely is better than being killed by one. Of course, if things go bad, it could be both ...
Re: (Score:1)
Durba klane psilo farras nulbo htoiler pfo nqu staka ksta. FO!
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
You don't need to -- they're now welcoming each other without any help from us.
Re: (Score:2)
Re: (Score:1)