Robot with a biological brain
John Shirley says: "Robot with a biological brain (wrote about this in my Eclipse novels, never thought it'd happen in my lifetime)."
The robot’s biological brain is made up of cultured neurons which are placed onto a multi electrode array (MEA). The MEA is a dish with approximately 60 electrodes which pick up the electrical signals generated by the cells. This is then used to drive the movement of the robot. Every time the robot nears an object, signals are directed to stimulate the brain by means of the electrodes. In response, the brain’s output is used to drive the wheels of the robot, left and right, so that it moves around in an attempt to avoid hitting objects. The robot has no additional control from a human or a computer, its sole means of control is from its own brain.A robot with a biological brain
UPDATE: Tom Simonite, online technology editor at Newscientist.com says: "I noticed you linked to the ZDnet coverage of robot with a rat's brain at Reading University. We've put together a video of it using its brain power to avoid obstacles."
the latest
latest episodes
So the article really means to say that they could have bribed Skynet with a bit of cheese?
What additional control could they have added?
>ts sl mns f cntrl s frm ts wn brn.
S, t's nt Rpblcn thn?
*rmsht*
I mean the neuronal response to the electrode is pre-determined, is it not? The neuron is not deciding, it's simply firing in response to stimuli...in an experimentally pre-determined way...it doesn't have a choice, this is "neuron as circuit element" IMO, nothing epistemological going on here.
Couldn't they use cloned heart cells in pumps too?
Reminds me of the "'Brain' in a dish flies flight simulator" http://www.cnn.com/2004/TECH/11/02/brain.dish/
Canuck--
what do you think your neurons are doing, if not "firing in response to stimuli"?
Neurons don't "decide" anything.
Does the robot have the strength of 5 gorillas?
Hey they do once you get enough of them working on each other.
Is this an unpredictable robot?
Here's a youtube video of the same robot:
http://www.youtube.com/watch?v=wACltn9QpCc
Everyone's seen the Technocalyps documentary, right?
Also, read the opening to the original Ghost in the Shell manga:
p.s. If anyone has one of those promotional posters of the first actual neurochip developed in 1980-something, I'd like to know where I can buy a print of it.
"Cultured neurons"... does that mean it's in a petri dish? What do they "feed" the "meat"?
I figure if you're going to make something with a meaty brain, you should probably figure out how to feed it. Otherwise it may die, subjecting you to a meaty subroutine called "regret". :)
Canuk, being meaty analog computers they would have several competing threshold curves affected by hormone, electrical, and nutrition levels as well as experience. It seems that they might also be capable of having a "bad day".
They didn't mention how a positive reward stimulus was delivered. Probably hormonal, but surprising in a way of thinking considering that this chip was able to in some sense be made "happy" when it was "good".
rebdav: Are these things subjective enough to qualify as moral agents, or beings with claims to moral rights? Or are we developing new slaves on purpose?
Sounds very similar to something I read back in 2000: http://www.abc.net.au/science/news/stories/s138960.htm -- Kevin Warwick gets a mention there as well. It gave me the willies at the time, especially since there was an almost-concurrent story about a robot that got its energy from digesting meat.
They say this story is new because there's no computer involved, and I kind of assume they know what they're talking about, although I don't see any mention of a computer in the earlier experiment.
To step away from the pure science of this for a minute: did anyone watch Star Trek: Voyager? This seems like a forerunner to the bio-neural gel packs they used as part of their computers.
Maybe they could hook it up with a voicebox, so it could let us know that it spends every single moment of its existence in a undiluted world of pure epic pain. By screaming.
nice one Mr. Voodoo...not like I was planning on getting any sleep tonight anyway.
can just imagine it banging itself against the wall to send the following morse message...
... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- . ... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- . ... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- . ... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- . ... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- . ... --- ... ... --- ... -.- .. .-.. .-.. -- . -.- .. .-.. .-.. -- .
Cordwainer Smith used Robots with Mouse Brains in several of his stories.
Well, I for one welcome our... oh, you know.
Really, Google, you don't have a built-in CW / Morse translator for your search field?
This doesn't make sense, as the article is written. There is no feedback. Sure, if the "sense" neurons are stimulated, and no "motor" neurons fire, the sense neurons will remain simulated (as the robot won't move away from the object). But so what? The "brain" doesn't care if the sense neurons are being simulated. It has no reason to learn anything about its environment, nor to learn to avoid obstacles or anything like that.
I don't doubt that the robot is moving, and moving in relation to objects, but this can be proved on paper. If neurons are stimulated, some will fire and some won't. You then take those firings and tell the wheels to move. Voilà, the robot moves in response to moving near objects.
But it's never going to change, except randomly. It's never going to learn to move away from objects, or to drive around them. It has no need to.
@zuzu http://morsecode.scphillips.com/cgi-bin/morse.cgi
Still no answer on how "sensate" ("sensable"?) something has to be to gain moral rights? Aren't these robots going to re-open the abortion debate in another context?
Spock mindmelds with the tiny thing, and tears run down his cheeks unabated.
"Jim.. Jim, he wants to leave this place."
I have to ask why post number 3 was disemvowelled. It seems to be an innocuous political joke. Are there any guidelines as to what is not acceptable, or is it all at the whim of the moderator? I don't mean to be impertinent, just want to know.
I think the complexity quotient will incorporate some type or measure of unpredictability of response to the same stimuli or sets of stimuli. Just a feeling...
It seems to be an innocuous political joke.
It has nothing to do with the post, it's offensive to many readers and it's not even funny.
Demon Seed!
Palindromic: ! Ha ha ha a ah ah ah !
Zuzu: Loving the references. *applause*
I wonder exactly how "smart" this thing is. I see from the article that they plan on measuring this by substituting machine brains and measuring it against those. I'd love to know how many neurons they managed to get on, how they compare to normal human neurons (what type they cultured, whether they grew comparably, etc).
Oh well, I guess we'll have to wait until they actually publish a paper. Seems reasonably exciting, though.
All it needs now is a steampunk case mod.
And a voice like Popeye's would add to its coolness.
Now that I've RTFA I'm very impressed...I can't believe no-one has yet mentioned the Borg-osity of it all, particularly the implants... but i still think it's cool.
You're right. And thus was born "pain".
Re: "I have no mouth and I must scream!"
Ever read Dalton Trumbo's "Johnny Got His Gun"?
Also was the basis for Metallica's 'One'.
Enjoyed a revival during early Vietnam-protest days. There was a movie....
great book, caused Trumbo no end of trouble. I rank it up there with All Quiet On The Western Front.
http://www.amazon.com/Home-Hangman-Roger-Zelazny/dp/1568652275
Supposed to be a new version this year.
Timothy Bottoms in 1971:
http://www.imdb.com/title/tt0067277/
This is... very exciting and intriguing.
Canuk, I felt bad about the rat feuses. The research is facinating, but this is a new ethics horizon. Since this robot in a sence might feel "pain" and "pleasure" what is that worth in the calculation of pain to a living thing. Scifi times we live in, where is my space car.
Zuzu is *so* on the money. Honestly, it continues to amaze me how the media always picks up on this fluff research, and even goes so far as to make outright lies in the write-up; like this one: "This new project is the first one to examine ‘how memories manifest themselves in the brain, and how a brain stores specific pieces of data.’". That's just false! It's totally giving the public the wrong impression.
This kind of stuff, like the "flight simulator" one mentioned above, doesn't actually teach us anything new. Most neuroscientists I know are simply not interested... It uses previously discovered neural phenomenon (synaptic plasticity and reward modultation) to build what amounts to a very expensive toy. It's a shame that so often science only becomes interesting to journalists when it involves meaningless bells and whistles...
The YouTube video mentions that the researchers "noticed" some neurons that responded to the sensor inputs "in a predictable way."
In other words, they found a couple neurons that simply echoed exactly what was signaled on their input electrodes. They used these neurons for two connection wires in a robot control circuit that the researchers designed, probably:
Something on the right --> (insert neuron here) --> back up right wheel.
Something on the left --> (insert neuron here) --> back up left wheel.
So, the neurons had no part in designing, evolving or modifying this circuit. They were just plugged into it. There was no learning, no reward or punishment.