How to teach a robot like a child ? | Maxime PETIT | TEDxÉcoleCentraleLyon

[Applause]

[Applause] [Music]

and everybody so I’m here to talk to you

and everybody so I’m here to talk to you about the specific field in robotics or

about the specific field in robotics or

about the specific field in robotics or developmental robotics can be summarized

developmental robotics can be summarized

developmental robotics can be summarized in a little simpler question how to

in a little simpler question how to

in a little simpler question how to teach a robot like a child at first I

teach a robot like a child at first I

teach a robot like a child at first I want you to think about a robot and I’m

want you to think about a robot and I’m

want you to think about a robot and I’m not talking about science-fiction robots

not talking about science-fiction robots

not talking about science-fiction robots I want to speak about a robot that

I want to speak about a robot that

I want to speak about a robot that already designed the world the ones that

already designed the world the ones that

already designed the world the ones that come to mind might actually be a

come to mind might actually be a

come to mind might actually be a specific kind of robot called industrial

specific kind of robot called industrial

specific kind of robot called industrial robots with big robotic arms walking

robots with big robotic arms walking

robots with big robotic arms walking into factories you can see them in

into factories you can see them in

into factories you can see them in action and I have another question

action and I have another question

action and I have another question well the humans they are nowhere to be

well the humans they are nowhere to be

well the humans they are nowhere to be found here because these robots are

found here because these robots are

found here because these robots are working in the factory in a specific

working in the factory in a specific

working in the factory in a specific setup where everything is known in

setup where everything is known in

setup where everything is known in advance and they are programmed in

advance and they are programmed in

advance and they are programmed in advance if you want to have robots in

advance if you want to have robots in

advance if you want to have robots in our home

our home

our home you know Ospital in order to help us

you know Ospital in order to help us

you know Ospital in order to help us with chores or Elpis to attend to people

with chores or Elpis to attend to people

with chores or Elpis to attend to people what do we do we are going to another

what do we do we are going to another

what do we do we are going to another kind of robotics called social robotics

kind of robotics called social robotics

kind of robotics called social robotics if it’s going social robotics is a

if it’s going social robotics is a

if it’s going social robotics is a specific field where actually the robot

specific field where actually the robot

specific field where actually the robot is among you is among human and yes to

is among you is among human and yes to

is among you is among human and yes to interact with them you can see some

interact with them you can see some

interact with them you can see some example here on your right you have the

example here on your right you have the

example here on your right you have the know for instance the null can help

know for instance the null can help

know for instance the null can help children in hospital children with

children in hospital children with

children in hospital children with diabetic condition to do dancing or

diabetic condition to do dancing or

diabetic condition to do dancing or exercising adding fat it can also be

exercising adding fat it can also be

exercising adding fat it can also be mounted on top of some wheelchair in

mounted on top of some wheelchair in

mounted on top of some wheelchair in order to act as copilot and one you if

order to act as copilot and one you if

order to act as copilot and one you if you have obstacles in front of you or

you have obstacles in front of you or

you have obstacles in front of you or behind you to the left you have another

behind you to the left you have another

behind you to the left you have another you man with robot called icon and will

you man with robot called icon and will

you man with robot called icon and will follow the trajectory in that case the

follow the trajectory in that case the

follow the trajectory in that case the robot would be a companion and you will

robot would be a companion and you will

robot would be a companion and you will be able to play with us some kind of

be able to play with us some kind of

be able to play with us some kind of different game musical game auto of

different game musical game auto of

different game musical game auto of annoyed in this specific setup you have

annoyed in this specific setup you have

annoyed in this specific setup you have one big passion for the robots is the

one big passion for the robots is the

one big passion for the robots is the adaptation to the environment why

adaptation to the environment why

adaptation to the environment why because we say we want robot in your own

because we say we want robot in your own

because we say we want robot in your own problem my own is not the same time

problem my own is not the same time

problem my own is not the same time yours if you want to put robot in the

yours if you want to put robot in the

yours if you want to put robot in the spit or your spittle in noodles not

spit or your spittle in noodles not

spit or your spittle in noodles not census battalion Paris so we don’t know

census battalion Paris so we don’t know

census battalion Paris so we don’t know where the robot will be when we could

where the robot will be when we could

where the robot will be when we could the other way is the robot will have to

the other way is the robot will have to

the other way is the robot will have to interact with the most complex object in

interact with the most complex object in

interact with the most complex object in the universe humans you will need to be

the universe humans you will need to be

the universe humans you will need to be able to communicate with it in order to

able to communicate with it in order to

able to communicate with it in order to add tense in the most efficiency way and

add tense in the most efficiency way and

add tense in the most efficiency way and then you will have to be multitask

then you will have to be multitask

then you will have to be multitask I want my robot to do my dishes to clean

I want my robot to do my dishes to clean

I want my robot to do my dishes to clean my dishes maybe you want them to do your

my dishes maybe you want them to do your

my dishes maybe you want them to do your troll I don’t know your laundry so we

troll I don’t know your laundry so we

troll I don’t know your laundry so we love to attend different tasks that we

love to attend different tasks that we

love to attend different tasks that we might know but maybe we don’t how do we

might know but maybe we don’t how do we

might know but maybe we don’t how do we deal with that we are taking a parody

deal with that we are taking a parody

deal with that we are taking a parody called development our robotics wise the

called development our robotics wise the

called development our robotics wise the idea is not to provide a robot that

idea is not to provide a robot that

idea is not to provide a robot that knows everything the idea is to provide

knows everything the idea is to provide

knows everything the idea is to provide the robots that can learn everything a

the robots that can learn everything a

the robots that can learn everything a little bit like a child provide the

little bit like a child provide the

little bit like a child provide the robot and if you need it to do something

robot and if you need it to do something

robot and if you need it to do something you just show you demonstrate it

you just show you demonstrate it

you just show you demonstrate it let the key challenge so let’s take a

let the key challenge so let’s take a

let the key challenge so let’s take a look at children development studies

look at children development studies

look at children development studies what make infant very good at learning

what make infant very good at learning

what make infant very good at learning about the wall at first you may think

about the wall at first you may think

about the wall at first you may think it’s because of the dexterity you know

it’s because of the dexterity you know

it’s because of the dexterity you know they are very good at grasping your

they are very good at grasping your

they are very good at grasping your smartphone and letting you go on the

smartphone and letting you go on the

smartphone and letting you go on the floor not to learn gravity but actually

floor not to learn gravity but actually

floor not to learn gravity but actually if you look at it and you compare it to

if you look at it and you compare it to

if you look at it and you compare it to our process you’re cruising so

our process you’re cruising so

our process you’re cruising so chimpanzee or Anto you want the same

chimpanzee or Anto you want the same

chimpanzee or Anto you want the same level more less on the dexterity level

level more less on the dexterity level

level more less on the dexterity level and the manipulative level so it might

and the manipulative level so it might

and the manipulative level so it might be needed but it’s not that it the key

be needed but it’s not that it the key

be needed but it’s not that it the key concept is in the social domain infant

concept is in the social domain infant

concept is in the social domain infant are very good to look at another person

are very good to look at another person

are very good to look at another person try to check where that gazing and

try to check where that gazing and

try to check where that gazing and focusing on that they are very good at

focusing on that they are very good at

focusing on that they are very good at trying to engage people to communicate

trying to engage people to communicate

trying to engage people to communicate with them and that’s very important

with them and that’s very important

with them and that’s very important because if you want to learn something

because if you want to learn something

because if you want to learn something you have to record the zone of proximal

you have to record the zone of proximal

you have to record the zone of proximal development of your gut ski it basically

development of your gut ski it basically

development of your gut ski it basically says that you can divide your task in

says that you can divide your task in

says that you can divide your task in three category your other one you can do

three category your other one you can do

three category your other one you can do on your own but actually you don’t earn

on your own but actually you don’t earn

on your own but actually you don’t earn a lot because you already know how to do

a lot because you already know how to do

a lot because you already know how to do it you are the task that you can’t do at

it you are the task that you can’t do at

it you are the task that you can’t do at all it’s not possible and you don’t

all it’s not possible and you don’t

all it’s not possible and you don’t stand so much because you don’t

stand so much because you don’t

stand so much because you don’t edge to do it the most interesting way

edge to do it the most interesting way

edge to do it the most interesting way the way you want to learn in order to

the way you want to learn in order to

the way you want to learn in order to optimize the learning rate is actually

optimize the learning rate is actually

optimize the learning rate is actually the zone of proximal development it’s

the zone of proximal development it’s

the zone of proximal development it’s always a task you can do on your own but

always a task you can do on your own but

always a task you can do on your own but you can do if somebody is helping you if

you can do if somebody is helping you if

you can do if somebody is helping you if your father of your mother is showing

your father of your mother is showing

your father of your mother is showing you is demonstrating something so that’s

you is demonstrating something so that’s

you is demonstrating something so that’s what we want and in order to do that we

what we want and in order to do that we

what we want and in order to do that we need to interact with our caretaker

need to interact with our caretaker

need to interact with our caretaker that’s what make human very unique among

that’s what make human very unique among

that’s what make human very unique among social and alien animal who are learning

social and alien animal who are learning

social and alien animal who are learning a language and you are able to

a language and you are able to

a language and you are able to communicate with other and the other are

communicate with other and the other are

communicate with other and the other are taking care of the children in order to

taking care of the children in order to

taking care of the children in order to doing that the children can then

doing that the children can then

doing that the children can then acquired capacity not from random

acquired capacity not from random

acquired capacity not from random movement but actually from learning from

movement but actually from learning from

movement but actually from learning from experts from their guidance what do we

experts from their guidance what do we

experts from their guidance what do we need if we want to build that inside the

need if we want to build that inside the

need if we want to build that inside the robot the first thing we need is a

robot the first thing we need is a

robot the first thing we need is a memory we need to be able to have the

memory we need to be able to have the

memory we need to be able to have the robot not forget everything if you

robot not forget everything if you

robot not forget everything if you switch out the robots and use every

switch out the robots and use every

switch out the robots and use every information every knowledge as to not be

information every knowledge as to not be

information every knowledge as to not be able to learn we need after to learn

able to learn we need after to learn

able to learn we need after to learn language in order to communicate with

language in order to communicate with

language in order to communicate with humans in order to learn different

humans in order to learn different

humans in order to learn different objects different concept like colors

objects different concept like colors

objects different concept like colors and to learn different action to do and

and to learn different action to do and

and to learn different action to do and then we need some kind of self reasoning

then we need some kind of self reasoning

then we need some kind of self reasoning or planning we need to be able to give

or planning we need to be able to give

or planning we need to be able to give to the robot capacity to use a knowledge

to the robot capacity to use a knowledge

to the robot capacity to use a knowledge also knowledge just learn and apply it

also knowledge just learn and apply it

also knowledge just learn and apply it to a problem or solution to create a

to a problem or solution to create a

to a problem or solution to create a sequence of action in order to good to

sequence of action in order to good to

sequence of action in order to good to go from front a to pond be too fast

go from front a to pond be too fast

go from front a to pond be too fast let’s go back to the neuroscience what

let’s go back to the neuroscience what

let’s go back to the neuroscience what is a long term autobiographical memories

is a long term autobiographical memories

is a long term autobiographical memories Neuman it’s the memories that make

Neuman it’s the memories that make

Neuman it’s the memories that make yourself a person it’s basically based

yourself a person it’s basically based

yourself a person it’s basically based on two different sub memory called the

on two different sub memory called the

on two different sub memory called the episodic memory it’s a memory about

episodic memory it’s a memory about

episodic memory it’s a memory about specific events that you can put in a

specific events that you can put in a

specific events that you can put in a specific place and time yesterday I saw

specific place and time yesterday I saw

specific place and time yesterday I saw a mouse in my kitchen

a mouse in my kitchen

a mouse in my kitchen that’s a pelagic memory and then you

that’s a pelagic memory and then you

that’s a pelagic memory and then you have the semantic memory

have the semantic memory

have the semantic memory it’s a memory about global knowledge and

it’s a memory about global knowledge and

it’s a memory about global knowledge and facts

facts

facts I know Mouse is a rodent and like cheese

I know Mouse is a rodent and like cheese

I know Mouse is a rodent and like cheese I don’t know where and when alone

I don’t know where and when alone

I don’t know where and when alone that’s but I know it that’s a semantic

that’s but I know it that’s a semantic

that’s but I know it that’s a semantic memory we need to build that in the

memory we need to build that in the

memory we need to build that in the robots the episodic memory will be a

robots the episodic memory will be a

robots the episodic memory will be a storing of all the data coming from the

storing of all the data coming from the

storing of all the data coming from the different sensor of the robots tactile

different sensor of the robots tactile

different sensor of the robots tactile sensor cameras joint actuators and the

sensor cameras joint actuators and the

sensor cameras joint actuators and the semantic memory will be concept at the

semantic memory will be concept at the

semantic memory will be concept at the robot Jastrow name of human has to

robot Jastrow name of human has to

robot Jastrow name of human has to interact we’ve set preferences here you

interact we’ve set preferences here you

interact we’ve set preferences here you can see the robot doing some movement

can see the robot doing some movement

can see the robot doing some movement and storing data inside the database you

and storing data inside the database you

and storing data inside the database you can solve different multi modal data so

can solve different multi modal data so

can solve different multi modal data so that camera here that you can see it’s

that camera here that you can see it’s

that camera here that you can see it’s also the John for instance and then we

also the John for instance and then we

also the John for instance and then we are able after to recall the specific

are able after to recall the specific

are able after to recall the specific events do you remember the last time

events do you remember the last time

events do you remember the last time yangjia showed you motor bubbling yes it

yangjia showed you motor bubbling yes it

yangjia showed you motor bubbling yes it was one month ago to be precise it was

was one month ago to be precise it was

was one month ago to be precise it was on January 26th actually I remember I am

on January 26th actually I remember I am

on January 26th actually I remember I am visualizing it right now you know that

visualizing it right now you know that

visualizing it right now you know that you have a memory let’s try to learn

you have a memory let’s try to learn

you have a memory let’s try to learn about language in order to do that we

about language in order to do that we

about language in order to do that we need first to learn the grammatical

need first to learn the grammatical

need first to learn the grammatical construction of the language we are

construction of the language we are

construction of the language we are building a brain with an artificial

building a brain with an artificial

building a brain with an artificial neural network and we’re giving as input

neural network and we’re giving as input

neural network and we’re giving as input sentence you put the toy on the left and

sentence you put the toy on the left and

sentence you put the toy on the left and you execute the action you provide also

you execute the action you provide also

you execute the action you provide also the meaning the output of the brain you

the meaning the output of the brain you

the meaning the output of the brain you do that several time and the brain will

do that several time and the brain will

do that several time and the brain will be able to go from one sentence to a

be able to go from one sentence to a

be able to go from one sentence to a specific meaning I will show you a video

specific meaning I will show you a video

specific meaning I will show you a video of this kind of results paying

of this kind of results paying

of this kind of results paying particular attention to the order of the

particular attention to the order of the

particular attention to the order of the world in the sentence and the order of

world in the sentence and the order of

world in the sentence and the order of the action to the sentence to execute or

the action to the sentence to execute or

the action to the sentence to execute or go between mode before you put the value

go between mode before you put the value

go between mode before you put the value on the left for the detail

on the left for the detail

on the left for the detail thinking of the situation

I have understood I think I have to

I have understood I think I have to point it on then I have to put my left I

point it on then I have to put my left I

point it on then I have to put my left I will point the guitar I will put the

will point the guitar I will put the

will point the guitar I will put the violin on the left

do you agree with this meaning yes so at

do you agree with this meaning yes so at the end you can see the robot he is

the end you can see the robot he is

the end you can see the robot he is asking the feedback from the human why

asking the feedback from the human why

asking the feedback from the human why because it’s not sure because it’s not

because it’s not sure because it’s not

because it’s not sure because it’s not sure about the solution the you man give

sure about the solution the you man give

sure about the solution the you man give a positive feedback you know that it

a positive feedback you know that it

a positive feedback you know that it learn well so it can use that you don’t

learn well so it can use that you don’t

learn well so it can use that you don’t not to learn even more so this running

not to learn even more so this running

not to learn even more so this running is not just one step in the room in the

is not just one step in the room in the

is not just one step in the room in the lifetime of the robot the learning will

lifetime of the robot the learning will

lifetime of the robot the learning will happen every time he’s doing something

happen every time he’s doing something

happen every time he’s doing something well now that we have a grammatical

well now that we have a grammatical

well now that we have a grammatical construction that’s fine but we seem to

construction that’s fine but we seem to

construction that’s fine but we seem to find the symbol behind it what is a toy

find the symbol behind it what is a toy

find the symbol behind it what is a toy what is a box what is left we need to

what is a box what is left we need to

what is a box what is left we need to create these symbols

create these symbols

create these symbols that’s called grounding symbols in order

that’s called grounding symbols in order

that’s called grounding symbols in order to do that both human and robots will go

to do that both human and robots will go

to do that both human and robots will go in some kind of joint attention they

in some kind of joint attention they

in some kind of joint attention they will both focus in a specific concept

will both focus in a specific concept

will both focus in a specific concept specific entity in the world and then so

specific entity in the world and then so

specific entity in the world and then so human will be able to provide the

human will be able to provide the

human will be able to provide the corresponding symbol the name of the

corresponding symbol the name of the

corresponding symbol the name of the object for instance and that I’m trained

object for instance and that I’m trained

object for instance and that I’m trained using pointing on one object in the

using pointing on one object in the

using pointing on one object in the scene

what can we learn we can learn a bit

what can we learn we can learn a bit object we can also learn about the body

object we can also learn about the body

object we can also learn about the body in order to control the ICAP watch plays

in order to control the ICAP watch plays

in order to control the ICAP watch plays alone out of my body

how do you follow this part of my body

how do you follow this part of my body this is yourself

this is yourself

this is yourself nice I know that I have a son change so

nice I know that I have a son change so

nice I know that I have a son change so here we just learn to the ICAP that he

here we just learn to the ICAP that he

here we just learn to the ICAP that he has hands with different finger and we

has hands with different finger and we

has hands with different finger and we name them so after we can use that in

name them so after we can use that in

name them so after we can use that in order to learn action that there would

order to learn action that there would

order to learn action that there would can do with the fingers doing under

can do with the fingers doing under

can do with the fingers doing under fright annotation I will my own finger

fright annotation I will my own finger

fright annotation I will my own finger can you describe the world actions and I

can you describe the world actions and I

can you describe the world actions and I will do you find yourself you read for

will do you find yourself you read for

will do you find yourself you read for yourself you feel yourself you are in

yourself you feel yourself you are in

yourself you feel yourself you are in for yourself so we have now very very

for yourself so we have now very very

for yourself so we have now very very simple action folding and unfolding

simple action folding and unfolding

simple action folding and unfolding fingers but we learn them from scratch

fingers but we learn them from scratch

fingers but we learn them from scratch so now what happened if we want to learn

so now what happened if we want to learn

so now what happened if we want to learn more complex action you can actually

more complex action you can actually

more complex action you can actually scare for them we are creating these

scare for them we are creating these

scare for them we are creating these blocks and we can create more complex

blocks and we can create more complex

blocks and we can create more complex blocks by assembling different one of

blocks by assembling different one of

blocks by assembling different one of them sequencing Z we are doing that by

them sequencing Z we are doing that by

them sequencing Z we are doing that by doing learning by instruction will

doing learning by instruction will

doing learning by instruction will explain the sequence to the robot for

explain the sequence to the robot for

explain the sequence to the robot for instance we can teach him how to close

instance we can teach him how to close

instance we can teach him how to close the hand which is basically folding all

the hand which is basically folding all

the hand which is basically folding all the fingers for yourself I follow my

the fingers for yourself I follow my

the fingers for yourself I follow my farm

fold your index I follow my edges fold

fold your index I follow my edges fold your meter I follow my middle fold your

your meter I follow my middle fold your

your meter I follow my middle fold your brain

brain

brain I followed my rain so here to fold the

I followed my rain so here to fold the

I followed my rain so here to fold the games finished sorry two for the hand of

games finished sorry two for the hand of

games finished sorry two for the hand of the iCub he actually has just to bend

the iCub he actually has just to bend

the iCub he actually has just to bend four fingers why because the fourth one

four fingers why because the fourth one

four fingers why because the fourth one and the fifth one actually tied with

and the fifth one actually tied with

and the fifth one actually tied with this robot the morphology of the robot

this robot the morphology of the robot

this robot the morphology of the robot is important when we teach something to

is important when we teach something to

is important when we teach something to it so this method can be applied to

it so this method can be applied to

it so this method can be applied to every robot that’s why it’s also

every robot that’s why it’s also

every robot that’s why it’s also difficult if you have different robot

difficult if you have different robot

difficult if you have different robot anyone to

anyone to

anyone to everybody part you have to do it for

everybody part you have to do it for

everybody part you have to do it for each robot each time is time-consuming

each robot each time is time-consuming

each robot each time is time-consuming after that you can also use this closing

after that you can also use this closing

after that you can also use this closing or opening end to create a ver complex

or opening end to create a ver complex

or opening end to create a ver complex action like crunching with your fingers

shoes three okay so right now why do we

shoes three okay so right now why do we have we have robots he has some kind of

have we have robots he has some kind of

have we have robots he has some kind of memory he learns the language he learned

memory he learns the language he learned

memory he learns the language he learned some action are we done

some action are we done

some action are we done no we don’t because the program is

no we don’t because the program is

no we don’t because the program is executing an action doing how to do an

executing an action doing how to do an

executing an action doing how to do an action is not enough you need to know if

action is not enough you need to know if

action is not enough you need to know if you can do it I can grab this butter I

you can do it I can grab this butter I

you can do it I can grab this butter I know I can but right now I can’t do it

know I can but right now I can’t do it

know I can but right now I can’t do it it’s too far away I have some

it’s too far away I have some

it’s too far away I have some precondition in grasping if you want to

precondition in grasping if you want to

precondition in grasping if you want to grasp an object the object should be in

grasp an object the object should be in

grasp an object the object should be in your reach and then you need to

your reach and then you need to

your reach and then you need to understand the effects of the action in

understand the effects of the action in

understand the effects of the action in the wall if I grasp a bottle at the end

the wall if I grasp a bottle at the end

the wall if I grasp a bottle at the end the bottle is not on the table anymore

the bottle is not on the table anymore

the bottle is not on the table anymore it’s in my hand and you need to

it’s in my hand and you need to

it’s in my hand and you need to understand this triptych precondition

understand this triptych precondition

understand this triptych precondition action the effect if you want to be able

action the effect if you want to be able

action the effect if you want to be able to plan and if you have a problem in the

to plan and if you have a problem in the

to plan and if you have a problem in the world you have a problem you want to

world you have a problem you want to

world you have a problem you want to reach a goal you need to have a sequence

reach a goal you need to have a sequence

reach a goal you need to have a sequence of action that is a link between the

of action that is a link between the

of action that is a link between the current state and the state I want to

current state and the state I want to

current state and the state I want to achieve and this sequence of action have

achieve and this sequence of action have

achieve and this sequence of action have to respect the precondition and that

to respect the precondition and that

to respect the precondition and that effect and the robot can do that by

effect and the robot can do that by

effect and the robot can do that by observing humans or observing the robot

observing humans or observing the robot

observing humans or observing the robot itself doing action and checking for

itself doing action and checking for

itself doing action and checking for regularities the regularities before an

regularities the regularities before an

regularities the regularities before an action will be the precondition the

action will be the precondition the

action will be the precondition the regularities after an action will be the

regularities after an action will be the

regularities after an action will be the effect so let’s say we want to teach the

effect so let’s say we want to teach the

effect so let’s say we want to teach the robot how to play a game to off an eye

robot how to play a game to off an eye

robot how to play a game to off an eye so you have to move different circle on

so you have to move different circle on

so you have to move different circle on different size and you have some rules

different size and you have some rules

different size and you have some rules for instance you can’t move the middle

for instance you can’t move the middle

for instance you can’t move the middle sized objects if you have the smaller

sized objects if you have the smaller

sized objects if you have the smaller object on top of it in the same way if

object on top of it in the same way if

object on top of it in the same way if you have a smaller object on the

you have a smaller object on the

you have a smaller object on the location you can put the middle you need

location you can put the middle you need

location you can put the middle you need to respect an order to create some kind

to respect an order to create some kind

to respect an order to create some kind of pyramid if you do that in front of

of pyramid if you do that in front of

of pyramid if you do that in front of the icon it will record all the data

the icon it will record all the data

the icon it will record all the data check the regularities and run the rule

check the regularities and run the rule

check the regularities and run the rule you will learn how to move the medium

you will learn how to move the medium

you will learn how to move the medium the big and the small object in too

the big and the small object in too

the big and the small object in too often oil is then you have a situation

often oil is then you have a situation

often oil is then you have a situation when you have all the object on the

when you have all the object on the

when you have all the object on the middle and you say I want all object to

middle and you say I want all object to

middle and you say I want all object to the left but

the left but

the left but pick the to our final groups and the

pick the to our final groups and the

pick the to our final groups and the robot will be able to compute a sequence

robot will be able to compute a sequence

robot will be able to compute a sequence of action in order to go from the states

of action in order to go from the states

of action in order to go from the states to the end by linking the precondition

to the end by linking the precondition

to the end by linking the precondition and the effect of action this sequence

and the effect of action this sequence

and the effect of action this sequence the specific sequence of action has not

the specific sequence of action has not

the specific sequence of action has not been shown to the robot the human just

been shown to the robot the human just

been shown to the robot the human just did some movement into alpha now but

did some movement into alpha now but

did some movement into alpha now but this solution is not known it’s computed

this solution is not known it’s computed

this solution is not known it’s computed by the robot

perfect goodbye so as conclusion if you

perfect goodbye so as conclusion if you don’t have social robots robots that

don’t have social robots robots that

don’t have social robots robots that will be able to interact with you we

will be able to interact with you we

will be able to interact with you we have taken a development approach

have taken a development approach

have taken a development approach meaning that we treat the robot like a

meaning that we treat the robot like a

meaning that we treat the robot like a child and we trying to teach him the

child and we trying to teach him the

child and we trying to teach him the robot has to learn and we are doing that

robot has to learn and we are doing that

robot has to learn and we are doing that by using child development theory we did

by using child development theory we did

by using child development theory we did some kind of VR equal learning with a

some kind of VR equal learning with a

some kind of VR equal learning with a human teacher the idea is to first read

human teacher the idea is to first read

human teacher the idea is to first read simple block and then to build on top of

simple block and then to build on top of

simple block and then to build on top of it to bid more complex ones

it to bid more complex ones

it to bid more complex ones then robot needs interaction with people

then robot needs interaction with people

then robot needs interaction with people there needs to be taken out of the lab

there needs to be taken out of the lab

there needs to be taken out of the lab put it into auspi Torah ohm in order to

put it into auspi Torah ohm in order to

put it into auspi Torah ohm in order to have data in order to learn about the

have data in order to learn about the

have data in order to learn about the world you confront the one is if you’re

world you confront the one is if you’re

world you confront the one is if you’re not in the war to win its large-scale

not in the war to win its large-scale

not in the war to win its large-scale and long term human robot interaction

and long term human robot interaction

and long term human robot interaction students in order to develop that we

students in order to develop that we

students in order to develop that we need to check what happened what the

need to check what happened what the

need to check what happened what the robot can learn if not by one day but

robot can learn if not by one day but

robot can learn if not by one day but for one week one month one year of

for one week one month one year of

for one week one month one year of training if you want your robots to go

training if you want your robots to go

training if you want your robots to go fetch you bear in your fridge and bring

fetch you bear in your fridge and bring

fetch you bear in your fridge and bring it back to you when you’re watching

it back to you when you’re watching

it back to you when you’re watching soccer or football we need that thank

soccer or football we need that thank

soccer or football we need that thank you for your attention and thank you for

you for your attention and thank you for

you for your attention and thank you for different laboratories that were part of

different laboratories that were part of

different laboratories that were part of his job or over Europe so young London

his job or over Europe so young London

his job or over Europe so young London Barcelona Sheffield London Italia Genoa

Barcelona Sheffield London Italia Genoa

Barcelona Sheffield London Italia Genoa thank

thank

thank [Applause]

[Applause]

[Applause] [Music]