Tuesday, May 26, 2015

Google plans can be dangerous for the world – Virtual Poland


  Lawrence E. Page, Google’s boss, he is a proponent of radical ideas, which some people consider to be the most dangerous for the world. As the ultimate goal puts on himself to create artificial intelligence.
 

 Menlo Park, California, 1999. Larry Page and Sergey Brin, founders of Google, sit happily between the damaged one dining table and workshop, where flickering CRT monitor rozbebeszone surrounded by computers, boxes of paper and tangled wires. In the half-dark garage can see the bike, it smells detergents and sweat. They call it the office. Here it is born a vision. – Artificial intelligence is the final version of Google – Larry Page said at that moment.
 

 Mountain View, California, 2014. The headquarters of futuristic appearance Page and Brin’s vision takes shape. Free access to information – from this we know Google. But Page and Brin wanting more: an intelligent technology that will help us in everyday life, encompassing global network. Techniques that will one day make us immortal. The search for artificial intelligence failed in 2005. Larry Page on California’s Mojave Desert, where the competition took place in the DARPA Grand Challenge. That’s what he saw, made a great impression on him.
 

 23 cars (so you can determine if a vehicle resembling a lawnmower or golf cart) which were autonomously – controlled entirely by computer – ride 212 km route studded obstacles. Most of the vehicles had not reached the finish line, she stopped after a few kilometers. One of the cars got stuck in a sand dune, another stuck in the bushes. In the next program … he rebelled and forced the vehicle to drive in a circle. Observers many times they had to evacuate from behind the barriers surrounding the route of the race.
 

  From the car to the artificial intelligence
 

 Only five vehicles arrived at their destination. Winner – Stanley VW Touareg – a seemingly normal car, if you miss the ambient scanner placed on its roof. During the race, the driver was sitting in the trunk, the vehicle headed the system using the Pentium M (1.6 GHz), which processed data from the laser scanner, camera and GPS module. Larry Page was so fascinated by the autonomous car that four years later he brought the vehicle designer Sebastian Thrun to Google. The headquarters of the company he started work on the project driving a car without a driver. They erected the first great step towards the realization of its vision.
 

 Autonomous cars initiated many bold projects: Google Brain supercomputer, humanoid robot, placed in balloons relay stations providing access to networks around the world. Google already is no longer the only search engine. Campus Company is today a real laboratory future. At this point you should ask an important question: do we have to enjoy with revolutionary products, or should we be afraid of this company with worldwide operations, managed by geniuses and driven by gigantic financial resources and computing?
 

 Made public purpose of the company is innocently: “Google’s mission to organize the world’s information to make it universally accessible and useful”. This message can be found on the home page of the company. – All the time we are still in the early stages – Larry Page said recently. – We’re working on this for 15 years, but still can not do that. Typing passwords to Google’s search engine using the keyboard is still too complicated, eg. On smartphones. – Control by voice is a much better way to reach information – Michiel Bacchiani said in an interview with CHIP-I. For Google working on speech recognition algorithm.
 

 It’s much easier it would be, according to Bacchianiniego, if a smartphone, taking into account the context, recognize what they are looking for (eg. At the airport: the road to board the aircraft), and there was no need to ask yourself this question. The intelligent system will provide information without being asked. – We discovered that our computers really do not know what we humans do and what we want – says Larry Page. Google to lack of intelligent search algorithm, which is based on fixed rules will process the data. Just how the system will be able to receive huge amounts of signals, instantly pick out relevant information and deduce the correct decision?
 

  Google’s computers learn to think
 

 Google mimics the human brain is a proven method using a neural network simulation. They reconstruct system of nerve cells and their interconnections (synapses). These connections carry information faster and keep it in any longer, the stronger the connection. Strengthening connections in humans it is called the learning process. A study is also capable of artificial neural networks. Assuming it took Google Deep Learning, simulated neural networks learn autonomously, without human intervention. Solutions based on artificial neural networks have been used for decades to simulate the brain.
 

 - Computing capabilities increased significantly today, as well as the amount of data that we have – said Michiel Bacchiani. – And this has greatly improved the capabilities of these models. As a result, it possible to reduce the error rate by about 25 percent. in the Android voice search using Deep Learning algorithm, eg. due to the fact that the background noise is better filtered. This is a big step for Google. – If you do thinks fit with Android digital assistant, speech recognition technology is the key – says Bacchiani.
 

  The first word of artificial intelligence Google: cat
 

 Mountain View, 2012. The overwhelming buzz of a thousand interconnected computers and their cooling devices signals: attention, mental work in progress! Powerful computer, unofficially called Google Brain, is ready to analyze ten million movies randomly selected from YouTube. Their content artificial brain has to sort on 22 000 established categories such as “human faces”, “cars”, “bicycles”. While performing this task 16 000 cores simulates the work of a billion nerve cells.
 

 On the monitor, scientists can see pictures of the face, houses and airplanes – is universally binding patterns for the category. The sorting process is proceeding according to plan. Suddenly, images of chaos emerging pattern that has not been defined: the cat’s mouth. Researchers are surprised – Google Brain alone learned to recognize cats. Simulated brain cells in the images were a pattern so often that he learned the artificial neuron responses at all the pictures of cats.
 

 Is it a proof of intelligence Google Brain? Replies can look at the research center at the Technical University in Munich. In the office of Professor neat and tidy. Only the robot’s arm suggests that performed here AI research. Prof. Alois Knoll does not share the euphoria investigators Google.
 

 - Standard cat emerges from the recognized image. However, the system does not understand that this is the cat – he says. Because understanding is the knowledge that the cat is an animal that can be raised and stroke, as well as the perception of humans as animals cat home. Google Brain found only mathematical patterns that enable the diagnosis of feline mouth in pictures. – Researchers are watching Google results say the system has recognized the cat – he said Knoll, who is involved in the Human Brain Project, an initiative to study and simulate the human brain. But in the end Google Brain fished only pattern that represents the class of a computer similar properties – it does not matter whether it is just a cat.
 

 
 Despite the enormous computing power of Google’s brain is missing one crucial aspect: – In robotics, we say the brain is not thinking – explains Dr. Florian Röhrbein, associate Knoll. And he adds: – He is here just to generate adequate behavior. And it can do only if it is in a body that is embedded in the environment.
 

 Intelligence therefore can not exist without a body. This conclusion is not lost as Larry Page’owi. It is therefore not surprising that in laboratories Google has long made the first attempts to take walking. He is responsible for Any Rubin, the creator of Android, in the 90s robotics engineer Carl Zeiiss. The benefit of AI systems is obvious for Google, but what he gains is to build robots?
 

 Officially, Google wants with the use of robots produce electronic parts and optimize freight logistics in warehouses. Although it is impossible to rule out the idea here repeatedly presented in a sci-fi novels: talking about intelligent, mechanical helper. It would be a logical consequence voice assistants. We can assume that it is also about Google to develop better prosthetic limbs. This technology in the future will help overcome human weaknesses and eliminate the shortcomings.
 

  – READ MORE –
 

  See also: The fastest bike in the world – a unique design
 

LikeTweet

No comments:

Post a Comment