The Resurrection of Einstein
When Einstein died, Goethe's poem was quoted to praise him-"He is like a dying comet, radiant, connecting infinite light with his light". Indeed, the light of Einstein, a scientific superstar, has been shining on the development of physics in the twentieth century. Therefore, exploring Einstein's way of thinking has always attracted people's interest. Some people even imagine whether it is possible to simulate Einstein's thinking law and let Einstein-type brain without Einstein continue to think about the latest physical problems.
What is Einstein's way of thinking? Einstein often meditated alone and immersed himself in his own thinking for hours on end; Sometimes, while boating and walking, he will suddenly take out his notebook, calculate something quickly, and then show a relaxed smile and magically make a simple explanation of complex problems. As his assistant Hoffman said, "This kind of magic is performed in Einstein's mind in a way that we can't guess." However, this does not mean that Einstein's thinking is unpredictable. In fact, we can not only analyze and trace the law of his thinking through his many papers and works, but also he has frankly disclosed his thinking characteristics in many works. For example, he pursues simplicity and tries to accommodate the most experimental facts with "the least hypothesis"; He often thinks through imagination and is good at "thinking experiments"; He admired aesthetic methods and thought that aesthetic concepts were one of the important factors leading to scientific development. Wait a minute. If we program Einstein's thinking method, can't we simulate Einstein's thinking?
Simulating the thinking methods of outstanding scientists with computers has become a research topic of many artificial intelligence laboratories in the world, and gratifying progress has been made. In recent years, Langley and others of Carnegie-Menon University in the United States put forward a Bacon program, which can make computers learn to use Bacon's induction method to summarize and reason a large number of experimental data, thus rediscovering many important laws discovered by some famous scientists in the 17 and 18 centuries, such as Kepler's third law, Ohm's law, Boyle's gas law and many other physical and chemical laws. Perhaps it won't be long before computers can quickly reproduce the scientific achievements made after centuries of humanization. A few years ago, the computer independently proved the famous four-color theorem; 1980, China mathematician Wu Wenjun proved Sim's theorem with a microcomputer.
This kind of simulated intelligent system can also transcend the individual limitations of thinking, gather the wisdom of many scientists, and form an "expert consultation system" that brainstorms and learns from others. Machine "think tank" is the most striking achievement in artificial intelligence research, which can be applied to military decision-making, economic planning, scientific research, enterprise management, medical diagnosis, traffic command, court trial and many other aspects. It is said that when the Nixon administration was in trouble in the Vietnam War, it turned to the machine "think tank" located in the Pentagon. After some calculation and planning by the "senior staff" of the machine, some schemes such as "carpet bombing" and "mine blocking the port" were put forward, which won some chips for the US government in the Paris peace talks. The "Intelligent Computer Application System of Traditional Chinese Medicine" designed by Li Taihang of Shanghai Institute of Computing Technology has concentrated the experience of many old Chinese medicine practitioners, and can learn and improve itself in diagnosis, with a preliminary level of intelligence.
Looking for the "soul" of wisdom
Someone asked: Can this intelligent system completely replace the human brain? It should be said that so far, although they have great advantages in computing speed and memory capacity, they lack flexibility and creativity, which is the "soul" of the human brain. Therefore, they are still inferior to the human brain. However, the research of artificial intelligence is to make progress in this respect. First of all, people try to make a "brain-like" intelligent machine by bionics, which is called "intelligent machine" or "cybernetic machine". Because the human brain is the most complex structure we know, there are 654.38+0.4 billion neurons in the cerebral cortex alone, and each neuron is connected with other neurons through tens of thousands of synapses, forming a complex neural network. Therefore, simulating the brain by simulating the specific structure is like "tigers eat the sky", but scientists have taken the first step. As early as 1943, American mathematical logician McCullough and neurophysiologist Pitts began to simulate the human brain from neurons, and now there are more than 100 neuron models. However, no matter how realistic the simulation of a single neuron is, it can't replace a brain. Any function of the brain is an integral feature organized by thousands of neurons in some way. So, on the basis of studying the structure of the cranial nerve, scientists simulated the function of the brain and made a "perceptron" that can classify graphics in 1958. 1969, Nakano of Japan proposed a "Lenovo machine" to simulate human association function. 1975, according to the new principle of synaptic mechanism of neurons, scientists put forward a neural network model of feedback cognitive machine. However, because the neural network system of human brain is extremely complex, it is as difficult to really understand its structure and mechanism, and to carry out mathematical description and electronic simulation as to build a huge and mysterious "cosmic city", so it still needs long-term and difficult exploration.
Besides, is there any other way? Some scientists try to bypass the giant peak of brain structure and use computer science and psychological methods to simulate the macro function of the brain. This method is to summarize the psychological activities such as steps, strategies and tricks used by people in solving various problems into laws, program them, and then try to simulate them with computers. For example, playing chess, solving problems and writing poems, people don't have to consider all possible solutions like computers, because people can first put forward assumptions based on experience for deliberation, which obviously narrows the range of choices. Simulating this thinking characteristic of human beings, a kind of "heuristic programming" appeared. This kind of computer problem solving is no longer based on the rigid algorithm program compiled in advance, but first makes some assumptions, then measures this assumption with practical arguments, and then puts forward new assumptions according to the results, and so on until a satisfactory solution is finally obtained. Someone designed a chess machine in this way. After ten to twenty hours of playing chess, the "chess player" can convincingly beat his programmer.
Structural simulation based on brain and functional simulation of computer software are two methods of artificial intelligence research at present. With the appearance of VLSI, when every neuron is likely to be simulated by an ultramicrocomputer on a silicon chip, these two research methods may be merged together to produce a real "computer".
Say goodbye to the "dinosaur era"
This subtitle seems strange. Dinosaurs are Mesozoic animals in geological history and have been extinct for more than 100 million years. Although some people have recently discovered that there seem to be traces of dinosaurs in the primitive jungle of Africa, we can't say that we are still in the dinosaur era, can we? In fact, this is just a metaphor for social development. Since primitive people lit the first bonfire, made the first stone tool and produced the first language to exchange ideas, the development of human society can not be separated from the three elements of energy, materials and information. It is precisely because of the evolution of materials that human society has experienced the Stone Age, Bronze Age, Iron Age, Steel Age and so on. Due to the continuous development of new energy and social productive forces, it led to the industrial revolution centered on steam engines in the 18th and 19th centuries. If human society is compared to a giant, then materials and energy are his flesh and blood. People can't live without flesh and blood, but only flesh and blood can't be a complete person. Only when there is intelligence to accept, transform, transmit and process information at the same time can the whole giant move and become an organic whole. Therefore, in the evolution and development of a society, bone, flesh and intelligence must develop harmoniously. Otherwise, it can only be regarded as a dinosaur at most-a huge body and a small intelligence. At present, human society has undergone several changes in matter and energy and made great progress. Only in information processing, we still haven't got rid of the limitations of human intelligence, as if we were in the "dinosaur era".
There is still a long way to go to bid farewell to the "dinosaur era". The rise of artificial intelligence will cause the information revolution of human society after the material and energy revolution. Some people call it the second scientific revolution. As long as you think about the great changes that have taken place in society after the first scientific revolution, you can imagine the bright future of the second scientific revolution. Looking forward to the future, one can't help but feel looking up at the radiant sun and staring at the starry universe. ...