Hod Lipson: Robots that are "self-aware"

Prevodilac: Mile Živković Lektor: Tatjana Jevdjic

So, where are the robots?

Pa, gde su roboti?

We've been told for 40 years already that they're coming soon.

Već 40 godina nam govore da roboti dolaze uskoro.

Very soon they'll be doing everything for us.

Uskoro će raditi sve za nas.

They'll be cooking, cleaning, buying things, shopping, building. But they aren't here.

Kuvaće, čistiće, ići u kupovinu, gradiće. Ali nisu ovde.

Meanwhile, we have illegal immigrants doing all the work,

Sa druge strane, ilegalni imigranti rade sav posao,

but we don't have any robots.

ali nemamo robote.

So what can we do about that? What can we say?

Šta možemo uraditi povodom toga? Šta možemo reći?

So I want to give a little bit of a different perspective

Želim da pružim drugačiju perspektivu

of how we can perhaps look at these things in a little bit of a different way.

kako bismo mogli da gledamo na ove stvari na drugačiji način.

And this is an x-ray picture

Ovo je rendgenski snimak

of a real beetle, and a Swiss watch, back from '88. You look at that --

prave bube i švajcarskog sata, iz 1988. Pogledajte to -

what was true then is certainly true today.

ono što je tad bilo istina i danas je.

We can still make the pieces. We can make the right pieces.

Još uvek možemo proizvesti delove. Prave delove.

We can make the circuitry of the right computational power,

Možemo napraviti kola prave računarske snage,

but we can't actually put them together to make something

ali ne možemo ih zapravo spojiti da bismo napravili nešto

that will actually work and be as adaptive as these systems.

što će zaista raditi i biti prilagodljivo poput ovih sistema.

So let's try to look at it from a different perspective.

Hajde da probamo da gledamo iz drugačije perspektive.

Let's summon the best designer, the mother of all designers.

Hajde da pozovemo najboljeg dizajnera, majku svih dizajnera.

Let's see what evolution can do for us.

Hajde da vidimo šta evolucija može da uradi za nas.

So we threw in -- we created a primordial soup

Napravili smo prvobitnu supu

with lots of pieces of robots -- with bars, with motors, with neurons.

sa puno delova robota - sa polugama, motorima i neuronima.

Put them all together, and put all this under kind of natural selection,

Sastavite ih zajedno i sve ovo prepustite mutaciji kao prirodnoj selekciji

under mutation, and rewarded things for how well they can move forward.

i nagradite stvari na osnovu toga koliko brzo se kreću unapred.

A very simple task, and it's interesting to see what kind of things came out of that.

Jednostavan zadatak i zanimljivo je videti šta je sve nastalo iz toga.

So if you look, you can see a lot of different machines

Ako pogledate, možete videti da je dosta različitih mašina

come out of this. They all move around.

nastalo iz ovoga. Sve se kreću.

They all crawl in different ways, and you can see on the right,

Pužu na različite načine i sa desne strane možete videti

that we actually made a couple of these things,

da smo zapravo napravili par ovih stvari

and they work in reality. These are not very fantastic robots,

i da rade u stvarnosti. Ovo nisu naročito fantastični roboti,

but they evolved to do exactly what we reward them for:

ali su se razvili da rade tačno onako kako ih nagrađujemo:

for moving forward. So that was all done in simulation,

za kretanje unapred. To je urađeno u simulaciji,

but we can also do that on a real machine.

ali možemo to uraditi i na pravoj mašini.

Here's a physical robot that we actually

Ovo je fizički robot kod kog zapravo

have a population of brains,

imamo populaciju mozgova,

competing, or evolving on the machine.

koji se takmiče ili se razvijaju na mašini.

It's like a rodeo show. They all get a ride on the machine,

To je kao rodeo šou. Svi mogu da jašu na mašini

and they get rewarded for how fast or how far

i nagrađuju se za to koliko brzo ili daleko

they can make the machine move forward.

mogu da pomere mašinu unapred.

And you can see these robots are not ready

Vidite da ovi roboti

to take over the world yet, but

još uvek nisu spremni da preuzmu svet,

they gradually learn how to move forward,

ali postepeno uče kako da se kreću unapred

and they do this autonomously.

i ovo rade samostalno.

So in these two examples, we had basically

Na ova dva primera, u principu smo imali

machines that learned how to walk in simulation,

mašine koje su u simulaciji naučile da hodaju

and also machines that learned how to walk in reality.

i mašine koje su naučile da hodaju u stvarnosti.

But I want to show you a different approach,

Ali želim da vam pokažem drugačiji pristup,

and this is this robot over here, which has four legs.

na ovom robotu, koji ima četiri noge.

It has eight motors, four on the knees and four on the hip.

Ima osam motora, četiri na kolenima i četiri na kukovima.

It has also two tilt sensors that tell the machine

Takođe ima dva senzora za nakretanje koji mu govore

which way it's tilting.

na koju stranu se nakreće.

But this machine doesn't know what it looks like.

Ali ova mašina ne zna kako to izgleda.

You look at it and you see it has four legs,

Vi je gledate i vidite da ima četiri noge,

the machine doesn't know if it's a snake, if it's a tree,

mašina ne zna da li je zmija ili drvo,

it doesn't have any idea what it looks like,

nema predstavu kako izgleda,

but it's going to try to find that out.

ali pokušaće da to sazna.

Initially, it does some random motion,

Isprva napravi nasumični pokret

and then it tries to figure out what it might look like.

i onda pokušava da shvati kako bi mogao da izgleda.

And you're seeing a lot of things passing through its minds,

Vidite puno stvari koje im proleću kroz um,

a lot of self-models that try to explain the relationship

puno samostalnih modela koji pokušavaju da objasne vezu

between actuation and sensing. It then tries to do

između aktiviranja i detekcije. Onda pokušava da uradi

a second action that creates the most disagreement

drugu radnju koja stvara najviše nesporazuma

among predictions of these alternative models,

između predviđanja ovih alternativnih modela,

like a scientist in a lab. Then it does that

poput naučnika u laboratoriji. Onda uradi to

and tries to explain that, and prune out its self-models.

i pokušava da objasni tu radnju i izdvoji model samog sebe.

This is the last cycle, and you can see it's pretty much

Ovo je poslednji ciklus i kao što možete videti

figured out what its self looks like. And once it has a self-model,

otprilike je shvatila kako izgleda. A kada ima model samog sebe,

it can use that to derive a pattern of locomotion.

može da ga iskoristi da smisli šablon kretanja.

So what you're seeing here are a couple of machines --

Ovde vidite nekoliko mašina -

a pattern of locomotion.

šablon kretanja.

We were hoping that it wass going to have a kind of evil, spidery walk,

Nadali smo se da će imati nekakav zloban hod, poput pauka,

but instead it created this pretty lame way of moving forward.

ali je umesto toga smislila ovaj prilično dosadan način kretanja unapred.

But when you look at that, you have to remember

Ali kad to pogledate, morate se setiti

that this machine did not do any physical trials on how to move forward,

da ova mašina nije fizički pokušala da se kreće unapred

nor did it have a model of itself.

i nije imala model svog izgleda.

It kind of figured out what it looks like, and how to move forward,

Otprilike je shvatila kako izgleda i kako da se kreće

and then actually tried that out.

i onda je to zapravo isprobala.

(Applause)

(Aplauz)

So, we'll move forward to a different idea.

Krenućemo ka drugačijoj ideji.

So that was what happened when we had a couple of --

To se desilo kada smo imali par -

that's what happened when you had a couple of -- OK, OK, OK --

to se desilo kada ste imali par - OK, OK, OK -

(Laughter)

(Smeh)

-- they don't like each other. So

- ne vole jedan drugog.

there's a different robot.

Dakle, to je drugačiji robot.

That's what happened when the robots actually

To se desi kada robote zaista nagradite

are rewarded for doing something.

za nešto što rade.

What happens if you don't reward them for anything, you just throw them in?

Šta se desi kada ih ne nagradite i samo ubacite unutra?

So we have these cubes, like the diagram showed here.

Imamo ove kocke, kao što vidite na dijagramu.

The cube can swivel, or flip on its side,

Kocka može da se rotira ili okrene na stranu,

and we just throw 1,000 of these cubes into a soup --

samo smo ubacili 1 000 ovih kocki u supu -

this is in simulation --and don't reward them for anything,

ovo je u simulaciji - i nismo ih nagradili ni za šta,

we just let them flip. We pump energy into this

samo smo ih pustili da se okreću. U ovo pumpamo energiju

and see what happens in a couple of mutations.

i gledamo šta će se desiti kroz nekoliko mutacija.

So, initially nothing happens, they're just flipping around there.

U početku se ne dešava ništa, samo se okreću u krug.

But after a very short while, you can see these blue things

Ali nakon kratkog vremena, možete videti

on the right there begin to take over.

da ove plave stvari sa desne strane počinju da preuzimaju vođstvo.

They begin to self-replicate. So in absence of any reward,

Počinju da se međusobno kopiraju. U nedostatku nagrade,

the intrinsic reward is self-replication.

suštinska korist je kopiranje samog sebe.

And we've actually built a couple of these,

Zapravo smo napravili par ovih robota

and this is part of a larger robot made out of these cubes.

i ovo je deo većeg robota napravljenog od ovih kocki.

It's an accelerated view, where you can see the robot actually

Ovo je ubrzan snimak gde možete videti da robot

carrying out some of its replication process.

izvršava deo procesa kopiranja.

So you're feeding it with more material -- cubes in this case --

Ako ih hranite sa više materijala - u ovom slučaju kockama -

and more energy, and it can make another robot.

i sa više energije, mogu da naprave još jednog robota.

So of course, this is a very crude machine,

Naravno ovo je veoma sirova mašina,

but we're working on a micro-scale version of these,

ali radimo na verziji mikroskopskih dimenzija,

and hopefully the cubes will be like a powder that you pour in.

s nadom da će kocke biti poput praška koji se može usuti

OK, so what can we learn? These robots are of course

Ok, šta možemo naučiti? Naravno, ovi roboti

not very useful in themselves, but they might teach us something

sami po sebi nisu veoma korisni, ali mogu nas naučiti nešto

about how we can build better robots,

o tome kako napraviti bolje robote

and perhaps how humans, animals, create self-models and learn.

i možda o tome kako ljudi i životinje prave modele sami sebe i uče.

And one of the things that I think is important

Jedna od bitnih stvari je

is that we have to get away from this idea

da moramo da zaboravimo na to

of designing the machines manually,

da dizajniramo mašine ručno

but actually let them evolve and learn, like children,

i da ih pustimo da se same razvijaju i uče, poput dece,

and perhaps that's the way we'll get there. Thank you.

i možda ćemo na taj način doći do toga. Hvala vam.

(Applause)

(Aplauz)