08:15
TED2007

Hod Lipson: Building "self-aware" robots

ホッド・リプソン「自己概念を持ったロボット」

Filmed:

ホッド・リプソンが、学習し、自己理解が可能で更に再生能力も有する、素晴らしいロボットを披露します。

- Roboticist
Hod Lipson works at the intersection of engineering and biology, studying robots and the way they "behave" and evolve. His work has exciting implications for design and manufacturing -- and serves as a window to understand our own behavior and evolution. Full bio

So, where are the robots?
さて ロボットはどこでしょう?
00:25
We've been told for 40 years already that they're coming soon.
ロボットがまもなく現れ仕事を代行してくれると
00:27
Very soon they'll be doing everything for us.
もう40年間ほど言われてきました
00:30
They'll be cooking, cleaning, buying things, shopping, building. But they aren't here.
料理 掃除 買物 建築 でもどこにも見当たりませんね
00:33
Meanwhile, we have illegal immigrants doing all the work,
代わりに違法移民が仕事をしていますが
00:38
but we don't have any robots.
ロボットはいません
00:42
So what can we do about that? What can we say?
さて どうしましょうか
00:44
So I want to give a little bit of a different perspective
そこで皆さんに今までとちょっと違った
00:48
of how we can perhaps look at these things in a little bit of a different way.
ロボットについての観点をお教えしたいと思います
00:52
And this is an x-ray picture
さて これはX線写真です
00:58
of a real beetle, and a Swiss watch, back from '88. You look at that --
本物のカブトムシと1988年スイス製の時計のものです
01:00
what was true then is certainly true today.
これらは当時からなんら変わっていません
01:05
We can still make the pieces. We can make the right pieces.
私たちは今なお正しい部品を作れますし
01:07
We can make the circuitry of the right computational power,
コンピュータ動力の電気回路も作れます
01:10
but we can't actually put them together to make something
しかし実はこれらを適応力があり実際に
01:13
that will actually work and be as adaptive as these systems.
動作する集合体としては組み立てられないのです
01:16
So let's try to look at it from a different perspective.
そこで異なる観点から見てみましょう
01:21
Let's summon the best designer, the mother of all designers.
究極のデザイナーを召喚しましょう
01:23
Let's see what evolution can do for us.
進化が私たちにもたらしたものを見てみましょう
01:27
So we threw in -- we created a primordial soup
我々は棒 モーター ニューロンといった
01:30
with lots of pieces of robots -- with bars, with motors, with neurons.
ロボットの材料をたくさん原生液に放り込みました
01:34
Put them all together, and put all this under kind of natural selection,
これらを混ぜ合わせ 自然淘汰と変異を経たのち
01:38
under mutation, and rewarded things for how well they can move forward.
どれだけ前進できるかによって報酬反応を与えたのです
01:42
A very simple task, and it's interesting to see what kind of things came out of that.
単純な課題ですが実験の成果を見るのはおもしろいですよ
01:46
So if you look, you can see a lot of different machines
この実験から生まれた異なる種類の機械を
01:52
come out of this. They all move around.
ご覧ください みな異なる方法で
01:55
They all crawl in different ways, and you can see on the right,
このように動き 這っています
01:57
that we actually made a couple of these things,
これらを実際に作ってみました
02:01
and they work in reality. These are not very fantastic robots,
素晴らしい見栄えではありませんが
02:03
but they evolved to do exactly what we reward them for:
前進に対する報酬反応を与えた分だけ
02:06
for moving forward. So that was all done in simulation,
シミュレーション上で 進化しました
02:10
but we can also do that on a real machine.
実際の機械を動かすことも出来ます
02:13
Here's a physical robot that we actually
ここに実際のロボットがありますね
02:15
have a population of brains,
機械上で競い合い 進化する
02:20
competing, or evolving on the machine.
頭脳部分を持っています
02:23
It's like a rodeo show. They all get a ride on the machine,
ロデオショーみたいでしょう
02:25
and they get rewarded for how fast or how far
機械がどれだけ速く 遠くへ前進できるかにより
02:28
they can make the machine move forward.
報酬が与えられます
02:31
And you can see these robots are not ready
ご覧のとおり このロボット達は
02:33
to take over the world yet, but
仕事を引き継ぐにはまだ未熟ですが
02:35
they gradually learn how to move forward,
次第に前進の仕方を学び
02:38
and they do this autonomously.
自律的に動くようになっています
02:40
So in these two examples, we had basically
さてこの2つの例を使って実際に
02:43
machines that learned how to walk in simulation,
シミュレーションで歩き方を学んだ機械と
02:47
and also machines that learned how to walk in reality.
現実に歩き方を学んだ機械を見ましたが
02:50
But I want to show you a different approach,
また別のアプローチをご覧頂きたいと思います
02:52
and this is this robot over here, which has four legs.
このロボットは 4本の脚を持ち
02:54
It has eight motors, four on the knees and four on the hip.
膝・腰それぞれに4つのモーターを搭載しています
03:00
It has also two tilt sensors that tell the machine
更に 2つのティルトセンサーを搭載し
03:02
which way it's tilting.
傾きを調べます
03:05
But this machine doesn't know what it looks like.
皆さんには4つの脚が見えますが 機械自身は
03:08
You look at it and you see it has four legs,
自分がどんな形なのか解ってません
03:10
the machine doesn't know if it's a snake, if it's a tree,
自分が蛇なのか木なのか
03:12
it doesn't have any idea what it looks like,
外見に対する知識は一切ありませんが
03:14
but it's going to try to find that out.
自らその特定を試みます
03:17
Initially, it does some random motion,
まず 適当に動いてみて
03:19
and then it tries to figure out what it might look like.
自らの形を探ろうとします
03:21
And you're seeing a lot of things passing through its minds,
いろいろなことを考えているんでしょうね
03:24
a lot of self-models that try to explain the relationship
動作と知覚の関係を説明しようとする
03:26
between actuation and sensing. It then tries to do
自己モデリングの試みです 次いで機械は
03:30
a second action that creates the most disagreement
第2の動きで予測とこれらの分析の
03:33
among predictions of these alternative models,
最大の不調和を引き出します
03:37
like a scientist in a lab. Then it does that
実験室の科学者のようですね
03:39
and tries to explain that, and prune out its self-models.
次にこの解釈に基づき 自己モデリングを絞り込みます
03:41
This is the last cycle, and you can see it's pretty much
この最後のサイクルで外見はかなり捉えられます
03:45
figured out what its self looks like. And once it has a self-model,
いったん自己モデリングを確立すると それを
03:48
it can use that to derive a pattern of locomotion.
運動パターン抽出のために利用します
03:52
So what you're seeing here are a couple of machines --
つまりご覧頂いているのは機械であり
03:56
a pattern of locomotion.
運動パターンなのです
03:58
We were hoping that it wass going to have a kind of evil, spidery walk,
蜘蛛の様に繊細で邪悪な歩き方を期待していましたが
04:00
but instead it created this pretty lame way of moving forward.
代わりにこのまどろっこしい前進法を生みました
04:04
But when you look at that, you have to remember
ところで思い出してください
04:08
that this machine did not do any physical trials on how to move forward,
この機械は前進の仕方を試行したことがありません
04:11
nor did it have a model of itself.
自己認識もありませんでした
04:17
It kind of figured out what it looks like, and how to move forward,
自らの形と前進法を何とか割り出して
04:19
and then actually tried that out.
実際にそれを実行したのです
04:22
(Applause)
(拍手)
04:26
So, we'll move forward to a different idea.
それでは また別のアイデアに移りましょう
04:31
So that was what happened when we had a couple of --
ご覧頂いたのは私たちがいくつか…
04:35
that's what happened when you had a couple of -- OK, OK, OK --
何が起こったかというと… OK わかった
04:40
(Laughter)
(笑)
04:44
-- they don't like each other. So
仲が良くないみたいですね
04:46
there's a different robot.
さてここに別のロボットがあります
04:48
That's what happened when the robots actually
先程ロボットの行動に対して報酬反応が
04:51
are rewarded for doing something.
与えられたらどうなるかを見ました
04:53
What happens if you don't reward them for anything, you just throw them in?
報酬反応を与えず 原生液に放り込めばどうでしょう
04:55
So we have these cubes, like the diagram showed here.
ここに図で示されているような立方体があります
04:58
The cube can swivel, or flip on its side,
この立体は面上で回転することができます
05:01
and we just throw 1,000 of these cubes into a soup --
ではこの立方体を1000個 原生液に入れてみましょう
05:04
this is in simulation --and don't reward them for anything,
報酬反応を無くした シミュレーションです
05:08
we just let them flip. We pump energy into this
動き回らせておき エネルギーを注入します
05:10
and see what happens in a couple of mutations.
変異を繰り返した後どうなるでしょう
05:13
So, initially nothing happens, they're just flipping around there.
最初は何も起きません ただ動き回っているだけです
05:16
But after a very short while, you can see these blue things
しかし少し経つと 右側の方で青色のものが
05:19
on the right there begin to take over.
増えているのが見えますね
05:23
They begin to self-replicate. So in absence of any reward,
増殖を始めたのです 外部からの報酬がなければ
05:25
the intrinsic reward is self-replication.
内因性刺激として 自己増殖を行います
05:29
And we've actually built a couple of these,
実際このようなものを造ってみました
05:32
and this is part of a larger robot made out of these cubes.
立方体から作られた大きなロボットの一部です
05:33
It's an accelerated view, where you can see the robot actually
早送りですが 実際にロボットが自己増殖を
05:37
carrying out some of its replication process.
行っているのがご覧いただけます
05:40
So you're feeding it with more material -- cubes in this case --
この場合は立方体ですが もっと部品とエネルギーが
05:42
and more energy, and it can make another robot.
あればまた別のロボットを作れます
05:46
So of course, this is a very crude machine,
もちろん これは大まかなものですが 我々は
05:49
but we're working on a micro-scale version of these,
ミクロの世界でも研究を行っています
05:52
and hopefully the cubes will be like a powder that you pour in.
そしてこの立方体を粒子程にしたいと考えています
05:54
OK, so what can we learn? These robots are of course
さてここから何が学べるでしょう これらのロボットは
05:57
not very useful in themselves, but they might teach us something
このままではあまり使い物になりませんが より高性能な
06:02
about how we can build better robots,
ロボットの作り方 もしくは人間や動物の
06:05
and perhaps how humans, animals, create self-models and learn.
自己モデリングの仕方に繋がるかもしれません
06:08
And one of the things that I think is important
そして大切なことの1つは
06:13
is that we have to get away from this idea
機械を人間の手で作り出すという
06:15
of designing the machines manually,
考えから離れ ロボットを子供のように
06:17
but actually let them evolve and learn, like children,
自由に進化 学習させるということ それが
06:19
and perhaps that's the way we'll get there. Thank you.
ロボットのある未来に繋がります ありがとうございました
06:22
(Applause)
(拍手)
06:24
Translated by Takahiro Shimpo
Reviewed by Junko Fundeis

▲Back to top

About the Speaker:

Hod Lipson - Roboticist
Hod Lipson works at the intersection of engineering and biology, studying robots and the way they "behave" and evolve. His work has exciting implications for design and manufacturing -- and serves as a window to understand our own behavior and evolution.

Why you should listen

To say that Hod Lipson and his team at Cornell build robots is not completely accurate: They may simply set out a pile of virtual robot parts, devise some rules for assembly, and see what the parts build themselves into. They've created robots that decide for themselves how they want to walk; robots that develop a sense of what they look like; even robots that can, through trial and error, construct other robots just like themselves.

Working across disciplines -- physics, computer science, math, biology and several flavors of engineer -- the team studies techniques for self-assembly and evolution that have great implications for fields such as micro-manufacturing -- allowing tiny pieces to assemble themselves at scales heretofore impossible -- and extreme custom manufacturing (in other words, 3-D printers for the home).

His lab's Outreach page is a funhouse of tools and instructions, including the amazing Golem@Home -- a self-assembling virtual robot who lives in your screensaver.

More profile about the speaker
Hod Lipson | Speaker | TED.com