- há 1 dia
Titulo Original: Documentário 2111 ROBÔS DO FUTURO Dublado HD
Canal Autor (Nome): Lukevi
Canal Autor (Link): https://www.youtube.com/@lukevi9250
Fonte do Video (Link): https://www.youtube.com/watch?v=2gpgCEC6Z1I
Licenca: Este conteudo e reutilizado sob a Licenca Creative Commons Atribuicao 4.0 Internacional (https://creativecommons.org/licenses/by/4.0/).
Note: The original content has not been modified. / O conteudo original foi mantido integralmente.
Canal Autor (Nome): Lukevi
Canal Autor (Link): https://www.youtube.com/@lukevi9250
Fonte do Video (Link): https://www.youtube.com/watch?v=2gpgCEC6Z1I
Licenca: Este conteudo e reutilizado sob a Licenca Creative Commons Atribuicao 4.0 Internacional (https://creativecommons.org/licenses/by/4.0/).
Note: The original content has not been modified. / O conteudo original foi mantido integralmente.
Categoria
📺
TVTranscrição
00:06In 2011 we will discover the latest advances in science and technology; let's go, we'll see what it will be like.
00:15Life in the next century will be truly incredible; could science create a new race?
00:23We'll go on vacation in space instead of superhumans, and our robots will help us with chores.
00:32everyday things, or will they one day have control of the earth? My name is Kevitt Index, and I invite you...
00:42you to watch with me at 2,111
01:11We may not have realized it, but robots already do it.
01:14A large part of our lives is spent in our homes, doing everyday tasks, and manufacturing our cars.
01:25And they even started warring over us; this could be considered a global conspiracy, the way in which...
01:34Artificial intelligence is already integrated into our world, and we depend on it.
01:43One day robots will evolve to the point where they will have their own minds and be like humans.
01:49The day will come when they will control the world.
01:54On this trip I will be accompanied by expert futurists who are dedicated to predicting how technology will affect us in the future.
02:03We don't realize how much artificial intelligence already drives our lives.
02:10civilization
02:13In 10, 20, or even 100 years, they will be a part of our lives in many, many ways.
02:20And no one will be surprised when they see a cyborg by their side.
02:40My first destination is Mexico City to see a competition where humans are still the best.
02:51Robots from around the world face off in a soccer match, a game that robotics engineers take very seriously.
02:57serious
02:58Scientists believe that in 2050 a team of humanoid robots will play against the reigning world cup champion and
03:06the robots will win
03:08One of the favorites this year is Dot Max, a team of 60-centimeter robots competing for the title.
03:16for the fifth consecutive time
03:20The team's human leader is Dr. Juan Manuel, and Barra is very confident.
03:29What do you consider a good robot?
03:32We consider a robot to be good when it can perform a minimum of functions, such as the movements of kicking a ball.
03:39Getting up and falling, walking and running, forwards, backwards and sideways.
03:47To execute such basic movements in soccer, the robot players rely on impressive technology.
03:56USB cameras for vision, a computer accelerometer and gyroscope for balance, and various servomotors for limb movements.
04:09They are autonomous robots; no one is controlling them; the computer they carry makes the decisions.
04:16He needs to feel, for example, what is causing the imbalance, equivalent to our inner ear, and he also needs to feel...
04:24the forces of the ground
04:26And most importantly, he needs to see, he needs to see the goal, where the ball is, his teammates and opponents.
04:36A job that even the worst player wouldn't find difficult, but for a robot it's a huge challenge.
04:44To help, the ball is always orange, the grass is green, and the lines are white.
04:50Even so, robots are still far from defeating humans.
04:53Barra believes this happens because we still don't know how to replicate the incredible capabilities of our brain.
05:01Neurons work in milliseconds, while computers work in nanoseconds.
05:06However, the human brain is more efficient.
05:09If I pick up a pencil and throw it to a colleague, they catch it, but a robot can't do that.
05:15And it's not just about mental capacity.
05:18Robots also need to improve their mechanics to compete with us.
05:23If one day they have a skeleton as efficient as ours—light, strong, and activated by muscles—they might just be able to do it.
05:33Then we'll have a robot similar to us.
05:45Although robots are very good at calculations, algorithms, and helping in factories
05:51The truth is, they're terrible at things that seem simple to us.
05:55how to walk
06:00The problem, the issue of transportation, is an extremely complicated one.
06:04It's not a trivial problem.
06:06So there are robots that can walk.
06:10But none of them can replicate the fluidity and type of control that we were able to achieve.
06:27Robots have not yet mastered the art of walking.
06:30But even so, they'll go far.
06:33with only 6 wheels
06:34The Curiosity rover explores the surface of Mars.
06:38but to master the more complex and irregular surface of the Earth
06:42The wheels are useless.
06:43Robots need to be able to move around as well as people do.
06:54My journey to investigate whether robots will one day walk alongside us.
06:58remains in the United States
07:00I'm going to Worcester Polytechnic Institute, which is outside Boston.
07:05to investigate a possible solution to the problem of walking
07:09I will meet Eduardo Torres.
07:11an engineer who believes the solution to the problem lies in more sensitive robots.
07:23Torres's enthusiasm for robots led him to leave his hometown of Cuenca in Ecuador.
07:28to follow his passion in Boston
07:32Why is it so difficult for robots to do something as simple for us as walking?
07:38unlike today's robots
07:40We have many sensors in our feet.
07:44So we have great sensitivity and can detect everything that exists on the ground.
07:50for Dr. Torres
07:51The secret to how to walk is not in having a more powerful brain.
07:55not even in a more precise view
07:56and yes, in sensitivity
07:58to prove his theory
08:00He proposes a simple test.
08:04Now I want you to understand that you are using information coming from the sensors in your foot.
08:10So you know where the edge of the step is.
08:12When can you take the next step?
08:14and is still aware of where his body weight is.
08:17and
08:18This is the same information we use when we are walking normally.
08:22It's just that it's unconscious for us.
08:27Our feet have more than 7,000 nerve endings.
08:30the visual information is incomplete
08:32And it can deceive us into believing that a thin layer of ice is a firm path.
08:38Feeling each step is more important than seeing.
08:43I consider the skin very important.
08:44based on the theory of artificial intelligence
08:47so-called embodied intelligence
08:49who argues that we are intelligent
08:51because we have a body
08:52And it is what allows us to feel and interact with the rest of the world.
08:58based on this theory
08:59Dr. Torres focused on making his robot's body more intelligent.
09:04Eduardo, but who is this robot?
09:06This is the hiker.
09:07And why do I only see half of it?
09:10because we are interested in the part that walks
09:13so we decided not to place so much emphasis on the top part.
09:17We believe there will always be a weight above.
09:21And that's enough.
09:22Tell me a little about what makes the hiker so special.
09:27he has a thumb
09:29which is important for walking, as you can see
09:33But perhaps the biggest difference with other current robots...
09:36whether these sensors are here
09:38touch sensors, the skin
09:42with heightened sensitivity
09:45It can detect where on the surface most of the weight is being applied.
09:49Thus, if the robot tilts to one side to take the next step
09:53The sensors indicate where the pressure is and what the next move should be.
09:59Wait, so that's basically what makes the walker a robot unlike any other in the world?
10:05That's right, it's the silicone rubber we use to create the structure and achieve the fingerprint effect.
10:12This allows us to measure normal and lateral forces.
10:15The normal ones here and the side ones that come like this
10:22The walker has 182 sensors.
10:25skin more like human skin
10:27and a computer capable of processing information
10:29Responding with quick and precise orders to avoid any setbacks.
10:35This is an important step for robotics.
10:39Little by little we are overcoming the challenge of making robots walk like all of us.
10:44But walking isn't the only skill that will be truly useful in the future.
10:48They need to learn to use their hands to work with us.
10:53There is already a robot that performs one of the most delicate manual tasks that exists.
10:58the surgeon's
11:09My journey in search of the robot most similar to humans.
11:13take me to Buenos Aires in Argentina
11:15where there is a robot that surpasses the skill of the human hand
11:18He's so good he was nicknamed Da Vinci.
11:25One of the specialists who works with him is Dr. Wenceslau Villamil.
11:29at the Italian hospital in the city of Buenos Aires
11:35The fact that we use robotic hands often allows us to reach places that are difficult to access.
11:43very small or distant vehicles with excellent visibility and excellent maneuverability.
11:48The robot has four arms.
11:51one for the cameras
11:52two that replicate the surgeon's movements
11:54and the fourth one who occupies the position of the surgeon's assistant.
11:59Good morning Letty, good morning Loren
12:02How are you?
12:03all good
12:03Is everything ready?
12:04all
12:04How are you?
12:06legal
12:06during surgery
12:08the surgeon in charge
12:10It doesn't need to be stereo.
12:12No need to wash or put on gloves
12:15And the robot's controls can even be located outside the operating room.
12:21in control
12:22The doctor sees through the Da Vinci robot's binocular system.
12:26which transmits images from two cameras
12:28one in each eye
12:29This allows the surgeon to view the operation as if they were inside the body.
12:34as if you were watching a 3D movie in a cinema.
12:38But the main feature of the Da Vinci robot is not the image.
12:42and yes, their robotic hands
12:44all the movements that a human being is capable of performing with their wrists
12:48are reproduced at the ends of the robotic instrument
12:52The surgeon has the option of having Da Vinci copy his movements.
12:56but on a smaller scale
12:57with a precision impossible for a human being.
13:01Furthermore, the robot's hands never tremble.
13:04which brings great benefits to the patient
13:06The robot makes small incisions.
13:09damages less tissue
13:11around the organs that we are going to treat
13:14and it also reduces the risk of bleeding.
13:17and blood transfusions
13:22The Da Vinci robot is not independent.
13:24and it still requires the surgeon's hand to operate.
13:26something that will certainly change in the future.
13:35Over time, we will program the robots themselves to be independent.
13:42It will happen, and it will be within the next 10 or 15 years.
13:46It will be very quick.
13:48humans who have to undergo surgery
13:51and they will choose more advanced procedures
13:55performed by surgical robots instead of humans
14:01I'm on a mission to discover inventions that will help our daily lives in the future.
14:07But I still haven't found that machine that solves everything I don't want to do.
14:14My next stop on my trip is Ithaca, in the United States.
14:18where a robot is learning to be a good homemaker
14:25here at Cornell University
14:27Yu Yang and his team work on Kodiak.
14:30a domestic robot that, when faced with learning algorithms
14:33You'll figure out how to manage in our house.
14:37Kodiak is the first personal assistant robot.
14:39We want a robot capable of performing all household chores.
14:46in many different places
14:47not specifically in an office, in a room
14:50Therefore, he must learn to make decisions on his own.
14:56Imagine you came home with a shopping bag.
15:00And he tries to help you put the food away in the refrigerator.
15:04First, he will look at the refrigerator.
15:06and identify the empty spaces within it.
15:10and then you'll be able to put things there
15:12So he's going to ask me for the object, which in this case is a bottle of beer.
15:17and will look in the refrigerator again for a space and a way to put the bottle.
15:25Our lives are full of implicit rules.
15:28From a young age we learn what each thing is for and what we can and cannot do.
15:33and get Kodiak to understand these rules
15:36It's hard work for Yu Yang.
15:38He is able to discern the shape and appearance of the object.
15:43For example, he can distinguish a martini glass from a bottle.
15:48It's a milk carton, based on volume and also appearance.
15:55That is, in color and texture.
15:57Our goal is to give the robot a better understanding of the object.
16:02For example, how can it be used?
16:05because Martini is drunk in that glass.
16:08And what can we drink here?
16:09These are concepts that are a bit more complex to understand than form or appearance.
16:16And that is our goal.
16:19Yu Yang believes that in a few years
16:22Robots like Kodiak will be in almost every home.
16:25A delicious dinner awaits us at the end of the day.
16:28so that we can enjoy the comfort of our homes.
16:32Today it is possible to have small domestic robots.
16:35who clean the floor while you go to work.
16:38so that when you return you will have a beautiful, clean house.
16:42It's clear that moving beyond this stage means becoming a personal robot.
16:47capable of cooking or performing more complex tasks
16:52It will take longer.
16:53But I definitely think it will happen in the future.
16:58the robot
17:03We want to create robots capable of imitating humans.
17:06to help humans
17:07and improve our quality of life
17:10We are learning.
17:11how to find the way
17:14of the new generation of human-imitating robots
17:19It seems our dream of having mechanics at home is becoming a reality.
17:22It's closer than we imagine.
17:25But robots can do more than just wash dishes.
17:29Actually, it's not such a distant future.
17:32There is a robot that works so well.
17:34that NASA has already hired him to work in space.
17:50My research aims to find out how far robots will go in 100 years.
17:55Now it takes off into space.
17:56NASA has developed what they call the world's most agile robot.
18:01the robonaut
18:04This humanoid robot was developed to assist astronauts in repair and maintenance tasks on spacecraft.
18:10It doesn't just look human.
18:12but it was designed as such
18:15with hands and arms
18:16This robot is capable of using the same tools as the crew.
18:27The Nauta robot is already working on the International Space Station.
18:31But for now, he'll just stay inside the spacecraft helping in the lab.
18:40But the next version will have the necessary features to withstand the extreme temperatures of space.
18:45It could even repair the exterior of spacecraft.
18:49instead of humans traveling to space
18:52We're talking about a space future.
18:55that will not be human
18:57Intelligent robots will be the ones traveling.
19:00I think it's more likely that they'll leave for those missions.
19:03in a way
19:05It will be easier to send a robot into space.
19:08for a 25-year mission
19:10than a human
19:11because robots can go further
19:13and to withstand more hostile environments
19:19We've already seen humanoid robots capable of walking.
19:22working in space
19:23or to perform cardiovascular surgery
19:25But not all scientists are inspired by humans to build them.
19:29Some are inspired by the animal kingdom.
19:34great scientists and important inventors
19:38They observed animals in search of inspiration.
19:41in the 15th century
19:43Leonardo da Vinci began studying the flight of birds.
19:46to create flying machine designs
19:48four centuries later
19:50We applied the same principles to build our first airplane.
20:00Now, this nature-inspired quest takes us to the outskirts of New York.
20:06John Long is the founder of the robotics research lab at Vassar College.
20:10He is a specialist in biomimicry.
20:13the study of nature-based technological projects
20:17one of the biggest challenges of biomimicry
20:19It's about understanding the world of animals and plants.
20:24which is very complex
20:25the simplest of animals
20:27It is more complex than anything built by a human.
20:30when it comes to developing robots
20:33Many engineers find inspiration in human beings.
20:36because their robots are designed to replace him.
20:39in everyday tasks of a terrestrial animal
20:41But in the oceans, we need to look for a different model.
20:46I'm not going to build a Michael Phelps robot.
20:49You know that swimming and exploring is for me.
20:51For that, I'll need a robot shark.
20:54thus the idea of combining biology and robotics
20:57becomes more powerful
21:01We are interested in a world we know little about.
21:04Today we know more about the surface of the moon than about the ocean floor.
21:09to improve the movements of their swimming robots
21:12Long studies the skeleton of sharks.
21:15since the spinal column of these animals
21:17It is a key component within its powerful propulsion system.
21:21the system that the professor wants to recreate
21:28This is where biomimicry comes in.
21:32We studied sharks in the pool.
21:34we study in the ocean
21:36And now we need to know more about the biomechanics of your skeleton.
21:40to create the artificial spinal columns that we call biomimetic spinal columns
21:46The spinal column mold will tell us something very important about how these animals swim.
21:51And our robots will have new abilities.
22:00We removed the shark's spine.
22:03and we conducted biomechanical tests to study its resistance
22:07How much energy does it store and what is its propulsion level?
22:10Now we need to figure out how to make an artificial spine to use in our robots.
22:19To create the replica, he uses a firm gelatin that mimics the consistency of shark cartilage.
22:33And now the robot will test its tail to see if it's efficient as a propulsion system.
22:49The goal is to test whether the artificial spine offers the expected propulsion for application in new robots that
22:57They will be able to move in the ocean.
22:59Mobility is not the only difficulty to be overcome in the development of an efficient aquatic robot.
23:05Positioning and communication systems on the seabed also pose challenges to underwater navigation.
23:15There are many laboratories around the world that are searching in the aquatic environment.
23:20Biologically inspired sensor systems to be incorporated into their robots for underwater navigation.
23:27An example of this is Festo, which draws inspiration from nature for intelligent communication between robots.
23:33Aqua penguins are the result of extensive research.
23:36These bionic penguins are able to navigate on their own thanks to a 3D sonar system.
23:43which, like in dolphins, allows them to communicate with their environment and with other robotic penguins.
23:53Sharks have an incredible ability to detect electric fields, and all animals are, in fact, electric.
24:00We generate small electrical signals with our muscles; sharks track other fish even without seeing them.
24:07And we want to understand this electromagnetic ability so that aquatic robots can navigate more efficiently.
24:15The shark possesses this sixth sense thanks to a network of electrosensory cells.
24:21which allows it to detect a signal as weak as the electricity generated by a muscle contraction.
24:27This also allows us to detect the Earth's magnetic field and use it for navigation.
24:33Engineers are very interested because they see ways to expand our knowledge through aquatic animals.
24:40And the reason for this is because they want new abilities for humans; they want abilities within the water.
24:47Many things we don't know about happen in the water, and we humans have to go there.
24:54For this, we need intelligent machines, such as biomimetic robots.
25:06We can draw inspiration from nature, but there's one thing that sets us apart from animals.
25:11Our intelligence, something that no one had been able to replicate until now.
25:23To understand how robots can adapt to our lives, my journey takes me to Japan.
25:40When we think of robots, we imagine machines that automatically repeat a task over and over again without rest.
25:47but they never reason on their own.
25:49This is about to change.
26:05Here at the University of Tokyo, they are programming a robot to learn gradually, which is known as learning.
26:12incremental
26:14Our robot has a special pattern neural network that we call soin.
26:21The acronym stands for Self-Organizing Incremental Neural Network, and it was designed specifically for online incremental learning.
26:30And all the time he learns new things.
26:32Yes, our robot can learn from both the internet and the real world, just like us.
26:39So these robots that perform household chores need this intelligence.
26:46The robot needs to adapt to the real world on its own.
26:54Dr. Hasegawa hopes that robots will be better prepared to face everyday problems.
27:00For example, preparing a cup of tea for the first time.
27:07This is a teacup; it's a type of cup, and here we're using a technical term.
27:17Transferable learning: the knowledge of how to make coffee is transferred with this new cup.
27:25And this knowledge is transferred to the new object.
27:30The robot needs to make a cup of green tea, but the only information in its system is how...
27:35make coffee
27:37He recognizes the concept of the cup and applies it to the task; he just needs to know what to find and how to prepare it.
27:43-lo
27:43then it connects to the internet; this robot has a system to select the information that will be used to complete its
27:50task
27:52This is the entrance.
27:54Imagine that these points represent information on the internet.
27:59So there are comments on Twitter, comments on Facebook, and things like that.
28:05There is a lot of information across this entire network.
28:10This is a noise zone.
28:12The Soin system filters out this noise, then it studies the shape and counts the number of groups.
28:23By organizing the information into groups, he can determine which is the most useful.
28:33The capabilities of the SOIN system can enhance the intelligence of many technologies.
28:44So what is the future of this type of robot?
28:48The SOIN system can be deployed in robots.
28:53also in cars, homes and in many devices
28:57So in the future we want to manufacture a brain.
29:03own artificial
29:04These artificial brains are already among us.
29:08and new technologies like soin
29:10This will lead to robots becoming increasingly intelligent and capable.
29:17When we talk about robots, what do we think of?
29:19We envision a technology capable of making decisions.
29:23your own decisions
29:25And this already exists around us.
29:28but in the case of fully independent robots
29:32capable of making decisions
29:34I believe we are very close to that.
29:37I believe that by 2030 they will be thinking for themselves.
29:41equal to or better than humans
29:44They will reason similarly to ours.
29:47but faster
29:48They will have better memory.
29:50combining the inherent advantages of today's computers
29:54perfect memory and high speed
29:56with the creative aspects
30:00subtleties of human thought
30:01in order to reproduce the subtlety
30:05and the complexity of the human brain
30:07Robotics engineers explore many paths.
30:10here in Japan
30:11another robot creator believes that the secret
30:14to achieve this great goal
30:15It's about starting with the smallest ones.
30:27My journey in search of the skills and future of robots continues in the city of Osaka, Japan.
30:33Here is one of the world leaders in robotics.
30:37Dr. Minoru Asada
30:39works on a revolutionary project
30:41which will change not only the robots
30:43but also our relationship with them
30:46Perhaps our expectations of robots are too high.
30:50We want them to start working as soon as we open the box, and to be smarter than us humans.
30:56Dr. Asada from Osaka University in Japan
30:59It was inspired by the way our children learn.
31:02We spent hours teaching them everything from how to walk to how to behave.
31:06And he thinks that if we do the same with robots...
31:09They will learn and behave more like us humans.
31:15to prove his theory
31:17Dr. Asada has a laboratory full of robots of different ages.
31:21He shows me the newest one.
31:23There are several robots.
31:25The first one is Neone.
31:27Its name comes from neonate or newborn.
31:32instead of the robot being pre-programmed
31:34learning from the internet
31:35Asada is developing a technique he calls mirror learning.
31:42So Neone sees a human do something and imitates it.
31:47That's right.
31:49we studied how the fetus
31:51the child or baby develops
31:54through interaction with their caregivers
31:57And we want to apply this model to robots in the future as well.
32:01then they will mirror us
32:03They will act like humans and perhaps think for themselves.
32:06and that?
32:09That's right, and that's one of the project's objectives.
32:11It's about developing robots that are similar to humans, like a mirror image or something like that.
32:16That's it, understand?
32:22Neone weighs only 3 kilos.
32:24It has touch sensors.
32:26cameras for vision
32:27microphones for listening
32:29processors and memory that allow it to learn
32:34How does a newborn begin life without knowing how to crawl or walk?
32:39He must learn everything from the humans around him.
32:42But once you learn it, you never forget it.
32:45Professor Asada's laboratory brings together different versions of the Neone robot.
32:50And everyone is learning from their human tutors.
32:55It's not exactly like our babies.
32:58But to humanize it, I'll teach you how to hug.
33:04First, I'm going to raise his arms.
33:12like this
33:16then I open his hands
33:18and below the arms
33:23ready
33:24We can ask Neone to repeat the movement he taught.
33:28to see if he understood correctly
33:30Can you do that for us?
33:35legal
33:38excellent
33:39And now Neone will remember the movement he taught forever.
33:43Wow, that's amazing!
33:45I think that's quite natural.
33:48It's very intuitive.
33:51It's... simply like a human baby.
33:55Dr. Asada also uses the method to teach robots to speak.
34:05These are just robots with voices.
34:07our
34:09Let's see.
34:11just look
34:12The caregiver takes charge of teaching.
34:15and the child, in addition to imitating the sound
34:19It also organizes Japanese vowels into categories such as
34:23A, I, U, E, U
34:25Yeah
34:30This type of model can also be used to explain
34:33How children learn Japanese vowels from their caregivers.
34:39In the end, the robot will learn everything it needs to deal with the world around it.
34:43Sure, if you have a caregiver with the patience to teach you.
34:49And what is the purpose of robots like these?
34:53This process can help people understand who the robots are.
34:58And this is important for understanding robots as part of society.
35:03Humans, society, must prepare for this.
35:08Otherwise, it will be difficult for robots to be accepted by society.
35:11So programming robots in this way will teach us a little more about ourselves.
35:16Yes, I think so.
35:21For Dr. Assada, the future involves more than just the evolution of robots.
35:25but also that of the human being
35:28In Japan, we believe in animism, that everything has a soul.
35:33For example, the box or the table, absolutely everything.
35:37Therefore, it is much easier for the Japanese to accept robots as part of society.
35:43Professor, how do you imagine robots will be in 100 years?
35:49we will have robots
35:51They will be very human.
35:54but they will not be treated like humans
35:58and yes, as companions
36:00or part of the family
36:02That is my wish.
36:07The interesting question is
36:09What will it be like when there are artificial robots?
36:11with emotions
36:12with sensitivity
36:14artificial sensitivity
36:15when they start to worry
36:17with them as a species
36:20and not just with humans
36:22Will this generate...
36:23another competition between forms
36:26of life on our planet
36:28particularly because
36:30Will they contain biological information?
36:32Yes, that's what awaits us.
36:35These robots are becoming increasingly agile and intelligent.
36:39They will undoubtedly be part of our future.
36:41But one question still remains.
36:44Will they ever evolve?
36:46To the point of taking control?
36:53my quest to find out to what extent robots
36:55They will be part of our lives.
36:57This ultimately leads me to Cornell University.
36:59in the United States
37:01A silent revolution is taking place.
37:04in Dr. Lipson's laboratory
37:06Your robots don't need us to say
37:08Or let's teach them what they have to do.
37:10They discover it for themselves.
37:17There are two schools.
37:19in one, the programmer or an engineer
37:21Think about all the things that could go wrong.
37:23in all possibilities
37:25and pre-program all these situations into the robot.
37:28the other option that nature chose
37:31It's about letting the robot learn on its own.
37:35But the robots don't have any computer code.
37:39that indicates how to move
37:40They should schedule it themselves.
37:46This here is one of the first self-programmable robots.
37:54This robot learned to walk on its own.
37:58But the interesting thing is that he doesn't know what he looks like.
38:02The robot receives the order to learn to walk.
38:05and just some information for that
38:09He doesn't know he has four legs.
38:12It only has eight motors and two spatial location sensors.
38:16It begins with random movements.
38:20gradually moves its motors randomly.
38:23He begins to feel it tilting to the sides and forward and backward.
38:29Through these movements, he gathers information and formulates hypotheses about himself.
38:36It might be a snake, a spider, a tree.
38:39and begins to discard certain hypotheses that do not correspond to reality.
38:44and then after about four days of doing that
38:48with these seemingly random movements
38:51he begins to form a self-image
38:54And from that image, he learns to walk.
38:58Lipson and his team practice the learning process.
39:01building robots that use the laboratory floor as their testing ground
39:06This robot is called Quadra Todd.
39:09The word Todd in English refers to Toddler.
39:12that is a developing baby
39:14Someone who knows nothing, doesn't know how to walk.
39:16Now we're going to show you the first steps.
39:20This here is the first generation, as you can see.
39:23She's in the first phase of learning to walk and doesn't walk very well.
39:27in the first, in the first group of steps, of walking patterns
39:31Perhaps one will move two centimeters and the other five.
39:35what walking five will
39:37probably seven, eight, and then ten, and so on.
39:41So this is the 40th generation of the simulated algorithm.
39:44We can see that it moves in a certain way.
39:48in a manner similar to a spider's walk
39:53or a four-legged animal
39:56Yes, we noticed that the legs are coordinated.
40:00He learned that it's helpful to move a few legs at a time.
40:03But even so, look, he can only move about ten centimeters from where he is.
40:11This is the 65th generation robot, and as you can see...
40:16It moves much more than the previous one.
40:18Just look, the legs have much more coordinated movements.
40:23Last time he moved about ten centimeters.
40:27and now he's reached about twenty or maybe thirty.
40:29And here we have the next robot from the 140th generation.
40:33You can see that he learned that it's better.
40:36move your body further off the ground.
40:38what can help in the movement
40:40The two legs are also much more coordinated.
40:43and he is able to move around better
40:48We lost a little control.
40:51by letting the machine learn on its own
40:55On the one hand, they learn more and go further.
40:59than if we program it in a specific way.
41:02On the other hand, we don't know exactly what they will learn.
41:05So we're waiting.
41:07And we'll see what happens.
41:09Rod Lipson and his team are working on the possibility of robots being programmable.
41:14and still project themselves
41:16a project that began in the virtual world
41:19We have millions of robots competing in a kind of game simulator.
41:23Here, the computer can build many robots with slight variations.
41:28test them, analyze their performance in the virtual world
41:32performing tasks, cooperating and competing with each other
41:35and then you can reproduce the ones that work best
41:38Duplicate them with slight variations and repeat the process.
41:42but the team discovered that the results with the simulators
41:45The results did not match those obtained with real robots.
41:49So the next step is to let the robots build themselves.
41:53using a 3D printer
41:55In our laboratory, we work with the idea of machines that build machines.
42:00that they build themselves, that they repair themselves.
42:04and adapt to new situations and improve over time.
42:07Lipson is beginning a process that may have implications for the robots of the future.
42:12once at that point
42:14Computers will be able to create their own designs.
42:19and they will become increasingly intelligent as they design their own next generation.
42:26maybe in 50, 100 or 200 years
42:28let's have machines with a level of intelligence
42:32similar to that of humans
42:34or perhaps superior to it in certain aspects
42:37It may be that the new generation of robots will come in the year 2030.
42:41manage to program yourself to feel emotions
42:44to be more intelligent than humans
42:47or to have abilities that are now unimaginable.
42:53On this journey I met robots that learn.
42:56robots that design and build themselves
42:58I don't know if robots will take over the world in 100 years.
43:02But I am certain that they will be an essential part of our lives.
43:06And they will become increasingly intelligent and stronger.
43:10But what I'm also sure of is that future engineers...
43:13They will always have an off button within easy reach.
43:16In this way, we humans will be in control.
Comentários