From the AI \u200b\u200bcrisis has created new areas of research IA such as
-
Neural Networks - Genetic Algorithms
- Distributed Artificial Intelligence
all have in common a new approach: intelligence is a product of the association, combination or connection of simpler independent entities (agents, genes or neurons) , which may have some and intelligence or any product of their interaction, the system would present an intelligent behavior. Which brings us to a new paradigm based on the copy of emerging systems and evolutionary nature, which in some ways is a return to the Cybernetics, the first technical science inspired by nature.
Research in the RNA have been immersed in building robots that learn to discover hidden relationships in data that can simultaneously encode and store information in a similar way as it does to the brain.
The Neurocomputation used as a basis the brain metaphor, but does not pursue the objective of building machines absolutely biologically plausible, but the development of useful machines.
intellectual capacity depends on the collective action of neurons that carry out processes in series and in parallel using feedback and a molecular and laminar organization with high capacity for self-organization and cooperativity based on a hierarchical structure that allows local processing centralized in phases.
Artificial Life is the name given to a new discipline was born in the 80, who studies the natural life by recreating biological phenomena in the computer and other artificial means, in order not only of theoretical understanding of phenomena under study, but also to discover and make practical and useful applications of biological principles in computer technology and engineering, for example, mobile robots, spacecraft, medicine, nanotechnology, industrial fabrication and assembly as well as other engineering projects.
In nature, evolution, particularly that of living beings, has certain characteristics that led to John Holland to start a research in an area that eventually became what today is called genetic algorithms (GA). The ability of a population of chromosomes to explore the search space "in parallel" and combine the best that it has been found through the mechanism of crossing-over (crossover), is intrinsic to the natural evolution and is being exploited by GAs.
From the biological standpoint, the problem centers on the imitation of the mechanism of evolution of living beings. Of a population are more likely to survive and have offspring those organisms best adapted to their environment. To combine two that have desirable characteristics for different aspects can arise both inherit new characteristics.
The premise of GAs, following the publication of the book of Holland "Adaptation in Natural and Artificial Systems "and the many researchers who use them as metaheuristic for optimization is that it can find approximate solutions to complex computational problems through a process of simulated evolution, in particular a mathematical algorithm implemented on a computer.
In Distributed Artificial Intelligence (DAI) studies collaborative problem solving by a group of distributed agents. Such cooperation is based on neither has the information (expertise, resources, etc) to fully resolve the problem and where an agent is characteristic of being a more or less autonomous entity with its own knowledge and environment and with the ability to interact with their environment and other agents.
Unlike expert systems, called by many "autistic systems" this new research dealing with that they are able to interact with the environment or what is like to be open and flexible.
Moreover, intelligent agents are emerging as one of the AI \u200b\u200bapplications more promising due to its close relation to the Internet (the third revolution in information). Without creating false expectations, we must remember the expert systems and intelligent agents given the advantages that account for network access. They are called to change our way of working to allow move users more comfortable and friendly environment. CONCLUSIONS
Until today, it is possible to develop systems that behave 'effective' in a specific subject, but not to 'emulate' human intelligence. The logic is only one edge of it: common sense, non-monotonic reasoning, recognition and other processes 'smart' are not yet fully resolved.
Chess programs are a tangible demonstration of what might be called an intelligence based on brute force, since their method is the computational speed, allowing you to search large areas and process a enormous amount of information processing speed that exceeds the human brain. Many believe that increasing the calculation speed of computers, which seems to have no limits, should lead to the emergence of so-called intelligent machines, a claim supported by the computers to be able to assess more variants of a problem, whether the game of chess , an estimate of engineering, industrial design or mechanical diagnosis, be able to achieve faster and more efficient solutions, relegating the human background. This argument is based on the accelerated growth that is evident in the computer hardware, while underestimating the development of software ignores the possibilities of the brain.
Today, science is no longer close in on herself and set in around him, understands that the best way to understand how something works in a human being is to understand first the simplest animals. So that science finds the emergent properties of intelligence, as a result of the complex interaction of simple elements and emulates the evolutionary genetic processes in the search for better solutions to really complex problems.
Artificial Intelligence 'conventional' in their attempts to mathematically model the processes of reasoning, could not achieve the dream of its creators, much less the IA followers called 'strong', those who have tried to convince us that thinking machines will be replaced. Begins a new era in Artificial Intelligence. The act of looking to nature will allow the man a step in the drive to achieve truly simulate human intelligence.
0 comments:
Post a Comment