What is thinking humanly in AI
What is thinking humanly in AI?
 |
| What is thinking humanly in AI |
Artificial Intelligence: Thinking Big | KQED QUEST
The following KQED production was produced with highly artificial intelligence, and this phrase may send a shiver down your spine, indicating that Terminator robot images are designed to dominate the world, but today's smart robots have advanced because they may be more like young children than trying to understand the world to show them Getting robots to recognize things from an image is one of those difficult things that robots have to do, so we often finish writing very large amounts of programs to say you know what a computer mouse is or what another object has to deal with complex tasks like recognizing On things AI researchers excel,
Like Andrew from Stanford University, in problems in cognitive psychology in computer science and even philosophy when trying to demonstrate intelligence in a robot we think about human intelligence and I think people can do a lot of things and when robots start doing some of the same things that we think require requires awareness or Mind-g or ideas This is what we think of robots designated to be intelligent and so for example if you have a robot that recognizes a new object that has never been seen before or if you are a robot capable of moving in a building there was an explicit program to navigate in these examples, what We think it is an AI, there was no F. A building robot that can recognize new objects and scale new environments, all in a one-day work to stare a Stanford AI artificial robot onto the stairs of the stairs project. A contract to develop technology to make it useful to put a robot in every home Many of us spend a lot of time today on housework like arranging, cleaning and cooking
And on the stairs project they want to build a robot they can do a lot of these things for you before the stairs, there was the first artificial intelligence robot created in 1966 at the Stanford Research Institute in Menlo Park since the shaky computers became smarter because I hadn't achieved similar gains in the form of Good with artificial intelligence, intelligence lies in studying hardware and more in the program and writing code or algorithms to make the machine more intelligent, the false computer algorithms were really at the heart of making the robot smart enough to perform all the tasks that we wanted to do to implement these algorithms It often requires programmers to have a deep understanding of both the principles of computer science as well as the principles of robotics about how things move and what different sensor readings tell you what is out there in the world and looking at the world through stairwells provides a new perspective that is more accurate and has a whole new dimension when You build a robot
You can often give him additional senses that as people we don't have one of the additional things he senses that we have on a robot a laser scanner and what it does is when it tries to know the free dimensional shape of something in front of it the laser plays what is in the father on that, and by seeing where it is located The laser beam and do some math can actually calculate its trivial position that has a lot of points in front of the robot but seeing the objects around it is only part of the stare the challenge faces, able to recognize it if it will interact with them, and so the way the robot recognizes Their things are that We don't actually show in many examples of saying bicycles and cups of coffee from a stapler or whatever the computer actually sees is a lot of numbers like this and it tells you about every pixel on that image exactly what is dark or gray, and so the small number is a light pixel and the big numbers It is a dark pixel and the task of the robot is to look at a lot of numbers like this and the number if there is a government exhibition of these numbers, and when it comes to learning, it has been shown that even machines can respond to some positive reinforcements here numbers work as rewards, so one way of thinkinMan Daphne Kohler of Stanford University uses artificial intelligence to make image searches more accurate by teaching computers to understand the context that the computer system is doing by itself, because it is largely stupid, it does not know clear things for you and me and everything that the computer system knows will be something that should You have to tell it or improve it and this is the revolution that we are going through and I have learned on its own by seeing enough examples from the real pictures, and here is an example that we use to identify giraffes in the beginning what is given to the system a set of specific giraffe examples for each of these ideas, there are in fact Algorithm for machine learning You learn the arrangement of different pixels in different areas of the image and say that this is a reasonable place for this purple dog, unlike this unreasonable matter, and the algorithm indicates that this thought in this giraffe corresponds to the same kind of position is this thought in this giraffe learns how to Making these differences like a human child Once he understands what giraffes look like, he can understand the difference between standing and relying on two or three examples that a computer system can do the same thing as artificial intelligence like teaching a child requires a lot of patience and flexibility, obviously That the algorithm committed a line
It's a very bad mistake if you really look closely you can try and understand the reason for making the mistake, there is a kind of contour line here that confuses a zebra and it looks like he sees a leg, and there are really two types of learning one can do using an example like this one is training the algorithm and the other It is our training for the person who designs the algorithm, but even if the machine can learn from errors, that real intelligence there will ever be a really smart machine that competed in its ability to think that the human mind is perhaps the most exciting computer we know in the entire universe, the brain contains Over ten to Eleven neurons, and this one is followed by eleven zero, and there are a large number, and today we hardly understand how the brain works but how the brain works and its structure may hold the key to a new world a kind of thinking machine just ask Dharmendra muda an IBM engineer at Almond Research Center in San Jose If you can draw the wiring diagram of the mind, then we can use the wiring diagram of the mind as a starting point for creating a new generation of computers: Dharmendra's way of re-engineering the human brain started on a slightly smaller scale One of my group's achievements was that we were able to assign a 4-brain brain to a computer An existing titan is what you are looking for here Ah rat brain, and represent the two tablets two upper half of the ball right, and represents the tables Ambasslaan hemisphere left, and this form of brain activity mice is an important step in their quest to bui if we imagine that there is a mind artificially, where treatments such as neurons and synapses may work
Imagine this visualization and basic arithmetic engine becoming a new paradigm in computer science, what kind of world is this Harold machine of good thinking doing in the near future? Not while the devices exist to perform such mental simulations, the program makes the cognitive cause of computers and the multi-tasking human brain remain out of reach today, but in the field of artificial intelligence, it is useful to think hugely and be patient. The things that we can spend our time knowing can be more exciting than trying to build systems, they are smart like you and you and whether it takes us 10 years or 50 years or a hundred years to know that I really don't know that you use them wisely, but I'm optimistic, I think One day, computers will be as smart as you, and I'm still free to find more and discover more at kqed.org slash. Search about that Unfortunately, the idea of rewards is distilled L. Mathematical formula, in which we use a number to tell the robot how to do it, and every time you dispense with a good thing, the reward function produces a large number and tells it that it has done well, the AI is about making robots smarter about getting computing applications like Those used on the Internet to think more cause like
Post a Comment