Android Copy of Young Woman Unveiled In Japan 264
An anonymous reader writes "According to IEEE Spectrum, Japanese roboticist Hiroshi Ishiguro, who had previously built a robot copy of himself, has now created a new android — and it's a 'she.' Geminoid F, a copy of a woman in her 20s with long dark hair, exhibits facial expressions more naturally than Ishiguro's previous android. 'Whereas the Geminoid HI-1 has some 50 actuators, the new Geminoid F has just 12. What's more, the HI-1 robot requires a large external box filled with compressors and valves. With Geminoid F, the researchers embedded air servo valves and an air servo control system into its body, so the android requires only a small external compressor.' It's also much better looking. Has the Japanese android master finally overcome the uncanny valley?"
uncanny valley (Score:1, Insightful)
we are still in it, it won't go away till we just can't tell at all, anything that looks slightly wrong will spark something in your brain to tell you all is not what it seems.
Just don't make it speak (Score:5, Insightful)
I'm down with my real doll android, aka Cherry 2000, being able to go get me a beer and thrust back, but as soon as she starts yappin at me, it's back to the whores.
Re:the question is... (Score:4, Insightful)
No. (Score:5, Insightful)
No.
Re:uncanny valley (Score:2, Insightful)
We're not even at the bottom of the valley yet. That thing is clearly a robot, nothing uncanny about it. Most importantly it needs a better mouth and small natural body movements.
Are you kidding? (Score:5, Insightful)
This is exactly the valley. It looks like a robot. The valley is huge, and it won't be solved by 12 actuators. Unless you're mostly blind.
Re:Dammit Japan. (Score:4, Insightful)
It's also a worthless neologism article [wikipedia.org] on wikipedia.
_What is it_ with people who dislike the documentation of neologisms on Wikipedia? Where else are you going to be able to find out what a word like "gynoid" means (which I'd never heard of before now)?
If anything, I'd favour the deletion of all non-neologised terms from Wikipedia on the grounds that nobody needs to know what a "table" is or read about Abraham Lincoln because they can always go out and buy a damn book.
The sheer irony of people calling for articles to be deleted because they are neologisms on the one hand, while praising Wikipedia on the other for being a fountain of contemporary knowledge on the other, is just beyond all understanding.
Why do we want human-looking robots now? (Score:5, Insightful)
Comment removed (Score:2, Insightful)
Re:Finally (Score:4, Insightful)
That's hardly a Troll.
Whoever produces the first practical, reasonably lifelike fuckbot is going to be very, very rich. Just because the rest of us don't want one doesn't mean they won't sell in droves.
After all, the inconvenient part about sex with another person is that it requires another person. A practical fuckbot would get rid of that barrier to entry. :)
But... (Score:2, Insightful)
...can she FOLD TOWELS?
Re:Dammit Japan. (Score:1, Insightful)
Re:Are you kidding? (Score:4, Insightful)
Seriously. The engineer is plainly making progress over his previous attempts, but he's got a long way to go. Humans are very good at reading subtle cues in other humans, like being able to tell when a friend is distracted by some unexpressed concern. That's what an android engineer is up against. This latest effort is immediately recognizable as a non-human when in motion, and something seems distinctly off about it in still pictures. And considering that even computer-rendered humans in movies -- which is arguably an easier problem -- are still less than 100% convincing in closeups, I'm not holding my breath for convincing androids any time soon.
hmm (Score:3, Insightful)
"exhibits facial expressions more naturally "
What facial expressions? All I saw in the video were mouth movement up and down and eyes moved.
The texture of the skin is amazingly natural as well as posture, but "facial expressions"?
Re:uncanny valley (Score:3, Insightful)
Re:Are you kidding? (Score:3, Insightful)
Re:No. (Score:3, Insightful)
IMHO "uncanny valley" just means "robotics engineers don't understand people". Whether we feel comfortable interacting with something has little to do with whether it visually approximates Homo Sapiens: Humans relate to other species all the time. Animated movies and TV shows are full of characters that don't resemble humans at all. A long scene [youtube.com] in the movie A.I. of an android grotesquerie that should epitomize the "uncanny valley" elicits sympathy, not fear. Why are these characters so easy to relate to? We empathize with them.
A fundamental ability of the brain is to relate to the minds of others. We call it "Theory of Mind", "empathy", or even "anthropomorphizing", though it's not a trait limited to humans. Social animals use it to maintain relationships. Predators and prey learn to predict each others' actions. On some level, our brain takes everything it percieves and tries to create a model to map its behavior to our own. The easier a thing's behavior is to understand, the more comfortable we feel. The harder a thing's behavior is to understand and predict, the more uneasy we become. Animators, storytellers, playwrights, and cartoonists learned long ago that the best way to make an audience comfortable with a character is not to create a convincing human but to create convincingly human mannerisms. They focus on the things humans look at in other humans: The shape of a mouth, the position of an eyebrow, a squint, the speed of a movement, the direction a gaze, a particular choice of words, an instinctive reaction; in other words, all the little things we subconsciously do to tell other humans what we are thinking. By replicating (and exaggerating) the mannerisms and behaviors of a human, you can portray a convincingly "human" character with a trash compactor and a pair of binoculars. [wordpress.com]
Unfortunately, the field of robotics hasn't caught up. Humans don't relate to today's androids, because they just don't have mannerisms we can relate to. Current "androids" are at best vaguely aware of the world, and uncoordinated in their interaction with it. If they are able to interact with humans at all they do so in a very limited manner, responding with incoherent, irrelevant, or parroted information, and their attempts at "emotion" or "human behavior" are artificial and hollow. We're unable to subconsciously translate their behaviors into our equivalents, as they are generated by algorithms that are structured very differently from our brains. Being with an android is more like being with an individual suffering some form of profound dementia.
Of course, there's no easy way to make robots interact fluently with us. Our abilities are too limited at the moment; we probably won't have "comfortable" androids until AI has taken a few huge leaps ahead. But all the uncanny valley really comes down to is that you can't simply make a robot that looks human without the mannerisms to go with it. In fact, the more organic a machine appears the harder it becomes to think of it as just a piece of equipment, so natural behavior becomes all the more critical.
Re:Are you kidding? (Score:2, Insightful)
The thing is, it's not even in the valley, it's nowhere near the valley. You look at it and think "Huh, it's a crappy animatronics doll", not "Hmm, that chick seems weird".
Re:The cost is terrifying. (Score:3, Insightful)
Yes, but it may make it easier to replace people at tasks where you don't need human-level sentience, like fast food. Just because we're safe for a little while in highly educated professions doesn't mean this can't profoundly change the world.
> Making a robot look and outwardly act like a human is a long way off from making one with human level sentience.
Re:Finally (Score:2, Insightful)
Re:Dammit Japan. (Score:2, Insightful)
urban dictionary?
google?
set up your own site if you think people need to know what the word is?
It is a big internet. Not everything needs to be on wikipedia. That isn't its purpose.
Re:Finally (Score:3, Insightful)
"* Does she run Linux?
Nope, Windows XXX.
- If not, can you install Linux with or without a 'hack'?
Yes, but installing anything from a floppy is obviously out of the question.
* What flavor of Linux would be best for a robot girlfriend? (and, no, you can't use 'GirLinux')
To be honest it doesn't much matter... the speech modules are poorly written no matter which you use, all of the personality packages clearly suffer from Asperger's, and the firmware for *ahem* down below was clearly written by somebody with no earthly idea of what it should be doing. Welcome to Linux.
* Will she still be considered 'female' once she's running Linux or will she be reset to a virgin teenage boy?
Gender differentiation is still in alpha... just expect androgyny and you won't be so disappointed.
* What type of DRM will she have?
Unauthorized users will be unable to open the legs.
- What will the pirate hack to break her DRM be called?
Roofies.
* And last, but not least, can she be programmed not to dump you?'
Who cares? She'll have a mute button anyway.
Re:uncanny valley (Score:3, Insightful)
It's not so much an feeling of "ooh, that's so weird" as going from the point where you're impressed at how well-made and interesting looking it is to being unimpressed at the remaining flaws.
In this case, it has a well-made face, but the fact that it doesn't move much except in starts and sputters makes it look palsied. It's not exceptionally creepy. It's just mildly unpleasant to look at in ways that something totally unrealistic like C3PO isn't.