Hod Lipson: Robots that are "self-aware"

Traducător: Tita Mihai Corector: Mirzac Iulian

So, where are the robots?

Deci, unde sunt robotii?

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

De 40 de ani ni se spune ca vor veni in curand.

Very soon they'll be doing everything for us.

In curand o sa faca totul in locul nostru:

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

o sa gateasca,o sa faca curat,o sa faca cumparaturi, o sa construiasca.Dar nu sunt aici.

Meanwhile, we have illegal immigrants doing all the work,

Intre timp ,avem imigranti ilegali care fac toata treaba,

but we don't have any robots.

dar nu avem nici un robot.

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

Asa ca ce putem sa facem? Ce putem sa spunem?

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

As vrea sa va arat o modalitate alternativa

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

despre cum ne putem uita la lucruri intr-un mod putin diferit.

And this is an x-ray picture

Aceasta este o radiografie

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

a unui gandac adevarat, si a unui ceas elvetian, din '88.Te uiti la --

what was true then is certainly true today.

ce era adevarat atunci cu siguranta este adevarat si astazi.

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

Inca mai putem sa fabricam piesele, putem face piesele corecte.

We can make the circuitry of the right computational power,

putem sa facem o placuta cu circuite de calcul,

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

dar nu putem sa le punem la un loc sa facem ceva anume

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

care sa functioneze cu adevarat si sa fie capabil sa se adapteze la fel ca aceste sisteme.

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

Asa ca sa incercam sa privim lucrurile dintr-o alta perspectiva.

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

Sa-l luam pe cel mai bun designer, cel mai bun designer dintre toti:

Let's see what evolution can do for us.

sa vedem ce poate face evolutia pentru noi.

So we threw in -- we created a primordial soup

Asa ca am amestecat-- am creat supa primordiala

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

cu multe bucati de roboti : cu fiare, cu motoare , cu neuroni.

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

Le adunam pe toate la un loc, si le supunem unui fel de proces natural de selectie,

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

unui proces de transformare, si vedem cat de bine au sa evolueze.

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

O sarcina foarte simpla, si e interesant de vazut ce fel de chestii rezulta.

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

Asa ca daca va uitati, o sa vedeti o gramada de masinarii diferite

come out of this. They all move around.

care au iesit din asta.Toate se misca,

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

intr-un fel sau altul,puteti vedea in dreapta,

that we actually made a couple of these things,

chiar am creat niste chestii de astea,

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

si chiar functioneaza. Nu sunt cine stie ce roboti,

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

dar au evoluat si au ajuns sa facea ce le-am cerut:

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

se mearga inainte.Toate aceastea au fost facute intr-o simulare,

but we can also do that on a real machine.

dar putem face asta si cu o masinarie reala.

Here's a physical robot that we actually

Acesta este un robot pe care avem

have a population of brains,

o populatie de creiere,

competing, or evolving on the machine.

care concureaza unele cu celelalte, sau evoluaza, pe robot.

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

E ca la un rodeo show: toti apuca sa controleze masinaria,

and they get rewarded for how fast or how far

si sunt recompensati pentru cat de repede sau cat de departe

they can make the machine move forward.

au facut masinaria sa mearga.

And you can see these robots are not ready

Dupa cum vedeti acesti roboti nu sunt gata inca

to take over the world yet, but

sa preia controlul asupra lumii,dar

they gradually learn how to move forward,

invata treptat cum sa se miste inainte,

and they do this autonomously.

si fac aceste lucru in mod autonom.

So in these two examples, we had basically

Deci in aceste doua exemple, am avut de fapt

machines that learned how to walk in simulation,

masinarii care au invatat cum sa mearga intr-o simulare,

and also machines that learned how to walk in reality.

si masinarii care au invatat sa mearga in realitate.

But I want to show you a different approach,

Dar vreau sa va arat o abordare diferita,

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

si acesta este robotul, aici, care are patru picioare,

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

are opt motoare , patru la genunchi si patru la solduri.

It has also two tilt sensors that tell the machine

Mai are si doi senzori care ii spun masinariei

which way it's tilting.

in ce parte sa se incline.

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

Dar aceasta masinarie nu stie cum arata.

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

Tu te uiti la ea si vezi ca are patru picioare,

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

masinaria nu stie daca e un sarpe, daca e un copac,

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

nu are nici o idee despre cum arata,

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

dar o sa incerce sa afle.

Initially, it does some random motion,

Initial, o sa incerce niste miscari aleatorii,

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

si apoi incearca sa afle cum arata --

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

si vedeti cum o gramada de lucruri ii trec prin minte,

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

o gramada de auto-modele care incearca sa explice relatia

between actuation and sensing. It then tries to do

dintre actiune si raspuns-- si apoi incearca

a second action that creates the most disagreement

o a doua actiune care creaza dezacord

among predictions of these alternative models,

printre predictiile modelelor alternative,

like a scientist in a lab. Then it does that

ca un om de stiinta intr-un laborator. Apoi face asta

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

si incearca sa explice, si sa isi intreaca concurentii.

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

Acesta e ultimul ciclu, si dupa cum puteti vedea

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

si-a dat seama cum arata,odata ce a avut un model dupa care sa se ia,

it can use that to derive a pattern of locomotion.

se poate lua dupa asta ca sa isi creeze un tipar de locomotie.

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

Deci ce vedeti aici este o adunatura de masinarii--

a pattern of locomotion.

un tipar de locomotie.

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

Speram ca o sa aiba un mers "smecher" ,ca al unui paianjen,

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

dar in schimb,si-a creat acest mod nasol de a se misca inspre inainte.

But when you look at that, you have to remember

Dar cand te uiti la asta , trebuie sa tii cont

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

ca aceasta masinarie nu stia cum sa se miste inainte,

nor did it have a model of itself.

nici nu stia cum arata.

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

Si-a dat seama cum arata , si cum sa se miste,

and then actually tried that out.

si apoi a facut o incercare.

(Applause)

(Aplauze)

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

Asa, se ne indreptam atentia spre o idee diferita.

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

Deci asta sa intamplat cand am avut o gramada de --

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

asta sa intamplat cand am avut o gramada de -- Ok ,Ok ,Ok--

(Laughter)

(Rasete)

-- they don't like each other. So

-- nu se plac.Deci

there's a different robot.

e un robot diferit.

That's what happened when the robots actually

Asta sa intamplat cand robotii

are rewarded for doing something.

au fost recompensati pentru ca fac ceva.

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

Dar ce se intampla cand nu ii recompensezi, doar ii arunci acolo?

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

Deci avem cuburile astea,dupa cum arata diagrama asta.

The cube can swivel, or flip on its side,

Cubul poate sa pivoteze ,sau sa sara pe o parte,

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

si aruncam 1,000 de cuburi de astea intr-o supa--

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

asta intr-o simulare-- si nu ii recompensam pentru nimic.

we just let them flip. We pump energy into this

ii lasam acolo sa sara. Le dam energie

and see what happens in a couple of mutations.

si vedem ce se intampla in cateva mutatii.

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

Initial, nimic nu se intampla, doar sar de colo colo.

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

Dar dupa o scurta perioada de timp,puteti vedea aceste chestii albastre

on the right there begin to take over.

din dreapta incep sa preia controlul.

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

Incep sa se auto-reproduca.Asa ca in absenta vreunei recompense,

the intrinsic reward is self-replication.

propria recompensa este auto-reproducerea.

And we've actually built a couple of these,

Si chiar am construit cativa din astia ,

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

si asta e o parte dintr-un robot mai mare facut din aceste cuburi,

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

e o filmare accelerata, in care puteti vedea cum robotul

carrying out some of its replication process.

urmeaza pasii spre procesul de replicare.

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

Deci o hranim cu mai mult material-- cuburi in cazul de fata--

and more energy, and it can make another robot.

si mai multa energie, si poate face un alt robot.

So of course, this is a very crude machine,

Dar desigur , aceasta este o masinarie foarte primitiva,

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

dar lucram la versiuni microscopice ale acestor masinarii,

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

si speram ca ,cuburile o sa fie ca o pudra pe care o adaugi.

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

OK, deci ce putem invatat? Acesti roboti nu sunt desigur

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

foarte folositori, dar ne pot invata cate ceva

about how we can build better robots,

despre cum putem sa contruim roboti mai buni,

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

si poate cum oameni , animalele, pot crea auto-modele si invata.

And one of the things that I think is important

Si unul din lucrurile pe care il consider important

is that we have to get away from this idea

este acela ca trebuie sa ne indepartam de idea

of designing the machines manually,

de a proiecta manual masinariile,

but actually let them evolve and learn, like children,

si in schimb sa le lasam sa evolueze si sa invete,precum copiii,

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

si poate in felul acesta o sa reusim . Multumesc.

(Applause)

(Aplauze)