TED2007
Hod Lipson: Building "self-aware" robots
Hod Lipson pravi "samosvesne" robote
Filmed:
Readability: 3.1
1,460,460 views
Hod Lipson pokazuje nekoliko svojih sjajnih malih robota, koji imaju sposobnost učenja, razumeju sami sebe i čak mogu i da kopiraju sami sebe.
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. Full bio
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
Double-click the English transcript below to play the video.
00:25
So, where are the robots?
0
0
2000
Pa, gde su roboti?
00:27
We've been told for 40 years already that they're coming soon.
1
2000
3000
Već 40 godina nam govore
da roboti dolaze uskoro.
da roboti dolaze uskoro.
00:30
Very soon they'll be doing everything for us.
2
5000
3000
Uskoro će raditi sve za nas.
00:33
They'll be cooking, cleaning, buying things, shopping, building. But they aren't here.
3
8000
5000
Kuvaće, čistiće, ići u kupovinu, gradiće.
Ali nisu ovde.
Ali nisu ovde.
00:38
Meanwhile, we have illegal immigrants doing all the work,
4
13000
4000
Sa druge strane, ilegalni imigranti
rade sav posao,
rade sav posao,
00:42
but we don't have any robots.
5
17000
2000
ali nemamo robote.
00:44
So what can we do about that? What can we say?
6
19000
4000
Šta možemo uraditi povodom toga?
Šta možemo reći?
Šta možemo reći?
00:48
So I want to give a little bit of a different perspective
7
23000
4000
Želim da pružim drugačiju perspektivu
00:52
of how we can perhaps look at these things in a little bit of a different way.
8
27000
6000
kako bismo mogli da gledamo na ove stvari
na drugačiji način.
na drugačiji način.
00:58
And this is an x-ray picture
9
33000
2000
Ovo je rendgenski snimak
01:00
of a real beetle, and a Swiss watch, back from '88. You look at that --
10
35000
5000
prave bube i švajcarskog sata, iz 1988.
Pogledajte to -
Pogledajte to -
01:05
what was true then is certainly true today.
11
40000
2000
ono što je tad bilo istina i danas je.
01:07
We can still make the pieces. We can make the right pieces.
12
42000
3000
Još uvek možemo proizvesti delove.
Prave delove.
Prave delove.
01:10
We can make the circuitry of the right computational power,
13
45000
3000
Možemo napraviti kola prave računarske snage,
01:13
but we can't actually put them together to make something
14
48000
3000
ali ne možemo ih zapravo spojiti
da bismo napravili nešto
da bismo napravili nešto
01:16
that will actually work and be as adaptive as these systems.
15
51000
5000
što će zaista raditi
i biti prilagodljivo poput ovih sistema.
i biti prilagodljivo poput ovih sistema.
01:21
So let's try to look at it from a different perspective.
16
56000
2000
Hajde da probamo da gledamo
iz drugačije perspektive.
iz drugačije perspektive.
01:23
Let's summon the best designer, the mother of all designers.
17
58000
4000
Hajde da pozovemo najboljeg dizajnera,
majku svih dizajnera.
majku svih dizajnera.
01:27
Let's see what evolution can do for us.
18
62000
3000
Hajde da vidimo šta evolucija
može da uradi za nas.
može da uradi za nas.
01:30
So we threw in -- we created a primordial soup
19
65000
4000
Napravili smo prvobitnu supu
01:34
with lots of pieces of robots -- with bars, with motors, with neurons.
20
69000
4000
sa puno delova robota -
sa polugama, motorima i neuronima.
sa polugama, motorima i neuronima.
01:38
Put them all together, and put all this under kind of natural selection,
21
73000
4000
Sastavite ih zajedno i sve ovo prepustite
mutaciji kao prirodnoj selekciji
mutaciji kao prirodnoj selekciji
01:42
under mutation, and rewarded things for how well they can move forward.
22
77000
4000
i nagradite stvari na osnovu toga
koliko brzo se kreću unapred.
koliko brzo se kreću unapred.
01:46
A very simple task, and it's interesting to see what kind of things came out of that.
23
81000
6000
Jednostavan zadatak i zanimljivo je videti
šta je sve nastalo iz toga.
šta je sve nastalo iz toga.
01:52
So if you look, you can see a lot of different machines
24
87000
3000
Ako pogledate, možete videti
da je dosta različitih mašina
da je dosta različitih mašina
01:55
come out of this. They all move around.
25
90000
2000
nastalo iz ovoga.
Sve se kreću.
Sve se kreću.
01:57
They all crawl in different ways, and you can see on the right,
26
92000
4000
Pužu na različite načine
i sa desne strane možete videti
i sa desne strane možete videti
02:01
that we actually made a couple of these things,
27
96000
2000
da smo zapravo napravili par ovih stvari
02:03
and they work in reality. These are not very fantastic robots,
28
98000
3000
i da rade u stvarnosti.
Ovo nisu naročito fantastični roboti,
Ovo nisu naročito fantastični roboti,
02:06
but they evolved to do exactly what we reward them for:
29
101000
4000
ali su se razvili da rade
tačno onako kako ih nagrađujemo:
tačno onako kako ih nagrađujemo:
02:10
for moving forward. So that was all done in simulation,
30
105000
3000
za kretanje unapred.
To je urađeno u simulaciji,
To je urađeno u simulaciji,
02:13
but we can also do that on a real machine.
31
108000
2000
ali možemo to uraditi i na pravoj mašini.
02:15
Here's a physical robot that we actually
32
110000
5000
Ovo je fizički robot kod kog zapravo
02:20
have a population of brains,
33
115000
3000
imamo populaciju mozgova,
02:23
competing, or evolving on the machine.
34
118000
2000
koji se takmiče
ili se razvijaju na mašini.
ili se razvijaju na mašini.
02:25
It's like a rodeo show. They all get a ride on the machine,
35
120000
3000
To je kao rodeo šou.
Svi mogu da jašu na mašini
Svi mogu da jašu na mašini
02:28
and they get rewarded for how fast or how far
36
123000
3000
i nagrađuju se za to
koliko brzo ili daleko
koliko brzo ili daleko
02:31
they can make the machine move forward.
37
126000
2000
mogu da pomere mašinu unapred.
02:33
And you can see these robots are not ready
38
128000
2000
Vidite da ovi roboti
02:35
to take over the world yet, but
39
130000
3000
još uvek nisu spremni da preuzmu svet,
02:38
they gradually learn how to move forward,
40
133000
2000
ali postepeno uče kako da se kreću unapred
02:40
and they do this autonomously.
41
135000
3000
i ovo rade samostalno.
02:43
So in these two examples, we had basically
42
138000
4000
Na ova dva primera, u principu smo imali
02:47
machines that learned how to walk in simulation,
43
142000
3000
mašine koje su u simulaciji naučile da hodaju
02:50
and also machines that learned how to walk in reality.
44
145000
2000
i mašine koje su naučile da hodaju u stvarnosti.
02:52
But I want to show you a different approach,
45
147000
2000
Ali želim da vam pokažem drugačiji pristup,
02:54
and this is this robot over here, which has four legs.
46
149000
6000
na ovom robotu, koji ima četiri noge.
03:00
It has eight motors, four on the knees and four on the hip.
47
155000
2000
Ima osam motora,
četiri na kolenima i četiri na kukovima.
četiri na kolenima i četiri na kukovima.
03:02
It has also two tilt sensors that tell the machine
48
157000
3000
Takođe ima dva senzora za nakretanje
koji mu govore
koji mu govore
03:05
which way it's tilting.
49
160000
3000
na koju stranu se nakreće.
03:08
But this machine doesn't know what it looks like.
50
163000
2000
Ali ova mašina ne zna kako to izgleda.
03:10
You look at it and you see it has four legs,
51
165000
2000
Vi je gledate i vidite da ima četiri noge,
03:12
the machine doesn't know if it's a snake, if it's a tree,
52
167000
2000
mašina ne zna da li je zmija ili drvo,
03:14
it doesn't have any idea what it looks like,
53
169000
3000
nema predstavu kako izgleda,
03:17
but it's going to try to find that out.
54
172000
2000
ali pokušaće da to sazna.
03:19
Initially, it does some random motion,
55
174000
2000
Isprva napravi nasumični pokret
03:21
and then it tries to figure out what it might look like.
56
176000
3000
i onda pokušava da shvati
kako bi mogao da izgleda.
kako bi mogao da izgleda.
03:24
And you're seeing a lot of things passing through its minds,
57
179000
2000
Vidite puno stvari koje im proleću kroz um,
03:26
a lot of self-models that try to explain the relationship
58
181000
4000
puno samostalnih modela
koji pokušavaju da objasne vezu
koji pokušavaju da objasne vezu
03:30
between actuation and sensing. It then tries to do
59
185000
3000
između aktiviranja i detekcije.
Onda pokušava da uradi
Onda pokušava da uradi
03:33
a second action that creates the most disagreement
60
188000
4000
drugu radnju koja stvara
najviše nesporazuma
najviše nesporazuma
03:37
among predictions of these alternative models,
61
192000
2000
između predviđanja
ovih alternativnih modela,
ovih alternativnih modela,
03:39
like a scientist in a lab. Then it does that
62
194000
2000
poput naučnika u laboratoriji.
Onda uradi to
Onda uradi to
03:41
and tries to explain that, and prune out its self-models.
63
196000
4000
i pokušava da objasni tu radnju
i izdvoji model samog sebe.
i izdvoji model samog sebe.
03:45
This is the last cycle, and you can see it's pretty much
64
200000
3000
Ovo je poslednji ciklus
i kao što možete videti
i kao što možete videti
03:48
figured out what its self looks like. And once it has a self-model,
65
203000
4000
otprilike je shvatila kako izgleda.
A kada ima model samog sebe,
A kada ima model samog sebe,
03:52
it can use that to derive a pattern of locomotion.
66
207000
4000
može da ga iskoristi
da smisli šablon kretanja.
da smisli šablon kretanja.
03:56
So what you're seeing here are a couple of machines --
67
211000
2000
Ovde vidite nekoliko mašina -
03:58
a pattern of locomotion.
68
213000
2000
šablon kretanja.
04:00
We were hoping that it wass going to have a kind of evil, spidery walk,
69
215000
4000
Nadali smo se da će imati nekakav zloban hod,
poput pauka,
poput pauka,
04:04
but instead it created this pretty lame way of moving forward.
70
219000
4000
ali je umesto toga smislila
ovaj prilično dosadan način kretanja unapred.
ovaj prilično dosadan način kretanja unapred.
04:08
But when you look at that, you have to remember
71
223000
3000
Ali kad to pogledate, morate se setiti
04:11
that this machine did not do any physical trials on how to move forward,
72
226000
6000
da ova mašina nije fizički pokušala
da se kreće unapred
da se kreće unapred
04:17
nor did it have a model of itself.
73
232000
2000
i nije imala model svog izgleda.
04:19
It kind of figured out what it looks like, and how to move forward,
74
234000
3000
Otprilike je shvatila kako izgleda
i kako da se kreće
i kako da se kreće
04:22
and then actually tried that out.
75
237000
4000
i onda je to zapravo isprobala.
04:26
(Applause)
76
241000
5000
(Aplauz)
04:31
So, we'll move forward to a different idea.
77
246000
4000
Krenućemo ka drugačijoj ideji.
04:35
So that was what happened when we had a couple of --
78
250000
5000
To se desilo kada smo imali par -
04:40
that's what happened when you had a couple of -- OK, OK, OK --
79
255000
4000
to se desilo kada ste imali par -
OK, OK, OK -
OK, OK, OK -
04:44
(Laughter)
80
259000
2000
(Smeh)
04:46
-- they don't like each other. So
81
261000
2000
- ne vole jedan drugog.
04:48
there's a different robot.
82
263000
3000
Dakle, to je drugačiji robot.
04:51
That's what happened when the robots actually
83
266000
2000
To se desi kada robote zaista nagradite
04:53
are rewarded for doing something.
84
268000
2000
za nešto što rade.
04:55
What happens if you don't reward them for anything, you just throw them in?
85
270000
3000
Šta se desi kada ih ne nagradite
i samo ubacite unutra?
i samo ubacite unutra?
04:58
So we have these cubes, like the diagram showed here.
86
273000
3000
Imamo ove kocke,
kao što vidite na dijagramu.
kao što vidite na dijagramu.
05:01
The cube can swivel, or flip on its side,
87
276000
2000
Kocka može da se rotira ili okrene na stranu,
05:04
and we just throw 1,000 of these cubes into a soup --
88
279000
4000
samo smo ubacili 1 000 ovih kocki u supu -
05:08
this is in simulation --and don't reward them for anything,
89
283000
2000
ovo je u simulaciji -
i nismo ih nagradili ni za šta,
i nismo ih nagradili ni za šta,
05:10
we just let them flip. We pump energy into this
90
285000
3000
samo smo ih pustili da se okreću.
U ovo pumpamo energiju
U ovo pumpamo energiju
05:13
and see what happens in a couple of mutations.
91
288000
3000
i gledamo šta će se desiti
kroz nekoliko mutacija.
kroz nekoliko mutacija.
05:16
So, initially nothing happens, they're just flipping around there.
92
291000
3000
U početku se ne dešava ništa,
samo se okreću u krug.
samo se okreću u krug.
05:19
But after a very short while, you can see these blue things
93
294000
4000
Ali nakon kratkog vremena,
možete videti
možete videti
05:23
on the right there begin to take over.
94
298000
2000
da ove plave stvari sa desne strane
počinju da preuzimaju vođstvo.
počinju da preuzimaju vođstvo.
05:25
They begin to self-replicate. So in absence of any reward,
95
300000
4000
Počinju da se međusobno kopiraju.
U nedostatku nagrade,
U nedostatku nagrade,
05:29
the intrinsic reward is self-replication.
96
304000
3000
suštinska korist je kopiranje samog sebe.
05:32
And we've actually built a couple of these,
97
307000
1000
Zapravo smo napravili par ovih robota
05:33
and this is part of a larger robot made out of these cubes.
98
308000
4000
i ovo je deo većeg robota
napravljenog od ovih kocki.
napravljenog od ovih kocki.
05:37
It's an accelerated view, where you can see the robot actually
99
312000
3000
Ovo je ubrzan snimak gde možete videti da robot
05:40
carrying out some of its replication process.
100
315000
2000
izvršava deo procesa kopiranja.
05:42
So you're feeding it with more material -- cubes in this case --
101
317000
4000
Ako ih hranite sa više materijala -
u ovom slučaju kockama -
u ovom slučaju kockama -
05:46
and more energy, and it can make another robot.
102
321000
3000
i sa više energije,
mogu da naprave još jednog robota.
mogu da naprave još jednog robota.
05:49
So of course, this is a very crude machine,
103
324000
3000
Naravno ovo je veoma sirova mašina,
05:52
but we're working on a micro-scale version of these,
104
327000
2000
ali radimo na verziji mikroskopskih dimenzija,
05:54
and hopefully the cubes will be like a powder that you pour in.
105
329000
3000
s nadom da će kocke biti
poput praška koji se može usuti
poput praška koji se može usuti
05:57
OK, so what can we learn? These robots are of course
106
332000
5000
Ok, šta možemo naučiti?
Naravno, ovi roboti
Naravno, ovi roboti
06:02
not very useful in themselves, but they might teach us something
107
337000
3000
sami po sebi nisu veoma korisni,
ali mogu nas naučiti nešto
ali mogu nas naučiti nešto
06:05
about how we can build better robots,
108
340000
3000
o tome kako napraviti bolje robote
06:08
and perhaps how humans, animals, create self-models and learn.
109
343000
5000
i možda o tome kako ljudi i životinje
prave modele sami sebe i uče.
prave modele sami sebe i uče.
06:13
And one of the things that I think is important
110
348000
2000
Jedna od bitnih stvari je
06:15
is that we have to get away from this idea
111
350000
2000
da moramo da zaboravimo na to
06:17
of designing the machines manually,
112
352000
2000
da dizajniramo mašine ručno
06:19
but actually let them evolve and learn, like children,
113
354000
3000
i da ih pustimo da se same razvijaju i uče,
poput dece,
poput dece,
06:22
and perhaps that's the way we'll get there. Thank you.
114
357000
2000
i možda ćemo na taj način doći do toga.
Hvala vam.
Hvala vam.
06:24
(Applause)
115
359000
2000
(Aplauz)
ABOUT THE SPEAKER
Hod Lipson - RoboticistHod 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.
Hod Lipson | Speaker | TED.com