The term “artificial intelligence” was created in 1956 upon the occasion of the important meeting at Dartmouth, where they met minds like Allen Newell, Herbert Simon, Marvin Minsky, Oliver Selfridge and John McCarthy. At the end of the 50 born the symbolic processing with a result of efforts Newell, Simon, and J. C. Shaw that instead of building systems based on numbers, engineered systems capable of manipulating symbols.
Thus, the different currents of thought in Artificial Intelligence, now called “Distributed Intelligence” have studied ways to establish, on the machines, “intelligent” behavior, which could be easily expressed by Minsky in the book Semantic Information Processing, “How do computers understand things? ”
The term ” Artificial Intelligence” can refer to an entire universe of programming techniques used to try to solve problems more efficiently than algorithmic solutions and the closest to the intelligent human behavior.
They stand out in the large family of techniques of artificial intelligence, the following research fields:
A) natural language, which addresses a set of techniques aimed at the recognition and generation of natural language, written and spoken. The main applications are in the field of universal translators, editors and the mining of texts and controls voice devices;
B) automation and robotics, i.e., the set of technological resources that aim to create autonomous robots, able to learn and make decisions; Many of these systems are already in operation and have been reported here in Hypescience.
C) perceptual systems, which aim to create visual pattern recognition systems, sound, and textures, to simulate and enhance the perception, whether optical, hearing or touch. Its main applications are in the medical diagnostic field and industrial quality control;
D) expert systems, which capture the knowledge in defined areas of knowledge and human experience, using it in decision-making. Observe currently important applications in medical diagnostics, identification of chemical compounds and in decision-making processes of business managers and brokers in the stock market. A special feature of expert systems is the occurrence of systems that support the decision in case-based reasoning;
E) genetic algorithms, which consist of several troubleshooting techniques based on the principles of Darwinian evolution, or mutation breeding and selection. They are used to solve problems involving a large number of variables and calculations, such as the fairings Aerodynamic projects developed by aerospace and automotive industry.
F) intelligent agents that characterize the set of stand-alone software that works in networks, or in parallel to a primary software created to achieve predictable, precise and repetitive tasks. For example, operating systems, software that manage e-mail and network tools are hosts for intelligent agents.
Noteworthy is its primary applications in the management of large volumes of information, such as, for example, in the stock market in search engines and the internet in the monitoring and management of e-commerce;
G) neural networks, which are simulations of the human brain processing patterns, such as plasticity and learning. Has its architecture based on an approximation of the animal brain and instead of being present, the neural network “learns” a particular training environment? Its construction is based on Perceptron, a discrete component that tends to simulate the physical behavior of a neuron. The association of thousands of perceptrons obtained enough plastic networks, which can recognize complex patterns, such as cracks in metal welds in pipelines or quality of an apple, diagnosed by color patterns of its bark.
In this book, we will talk about artificial intelligence from the time that we would never think that such a concept existed.