http://www.crete-holiday.org/crete-greece-ghost/
Artificial Intelligence
Artificial intelligence (AI) is the intelligence of machines and the branch of computer science that aims to create it. Textbooks define the field as "the study and design of intelligent agents" where an intelligent agent is a system that perceives its environment and takes actions that maximize their chances of success. John McCarthy, who coined the term in 1956, defines it as "the science and engineering of making machines intelligent. "
The camp was founded on the assertion that a fundamental property of human beings, intelligence-the wisdom of Homo sapiens can be so precisely described that can be simulated by a machine. This raises philosophical questions about the nature of mind and the limits of scientific arrogance, issues have been addressed by the myth, fiction and philosophy since antiquity. Artificial intelligence has been impressive optimism, has suffered setbacks spectacular and today has become an essential part of the technology industry, providing heavy burden for many of the most difficult problems in science computer.
AI research is highly skilled, deeply divided into subfields, which often fail to communicate. Subfields have grown up around particular institutions, the work of individual researchers, the solution of specific problems, the longstanding differences of opinion about how the AI should be done and the implementation of the wide variety of tools. The central problems of AI include features such as reasoning, knowledge, planning, learning, communication, perception and the ability to move and manipulate objects. General intelligence (or "strong AI") is still a long-term (investigate).
History
Thinking machines and artificial beings appear in Greek myths, such as Talos Crete, the golden robots of Hephaestus and Pygmalion's Galatea. Human similarities believes that intelligence is built into all major civilizations: animated statues are worshiped in Egypt and Greece and humanoid robots were built by Shi Yan, Hero of Alexandria, Al-Jazari and Wolfgang von Kempelen. It is also believed that artificial beings have been created by J? Bir ibn Hayy? No, Judah Loew and Paracelsus. In the 19th and 20th centuries, artificial beings has become a common feature in fiction, as in Mary Shelley, Frankenstein or Karel? Universal APEK RUR (Rossum's Robots). McCorduck Pamela contends that these are all examples of an ancient impulse, as she describes it, "to forge the gods." The stories of these creatures and their fates to examine many of the same hopes, fears and ethical concerns presented by artificial intelligence.
Mechanical or "formal" reasoning has been developed by philosophers and mathematicians since antiquity. The study of logic led directly to the invention of electronic programmable digital computer, based on the work of mathematician Alan Turing and others. The theory Turing suggested that a computing machine, shuffle through symbols as simple as "0" and "1" could simulate any act conceivable mathematical deduction. This, together with recent findings of neuroscience, information theory and cybernetics, inspired by a small researchers to begin to seriously consider the possibility of building an electronic brain.
The field of AI research was founded at a conference on the campus of Dartmouth College in the summer of 1956. The guests, including John McCarthy, Marvin Minsky, Allen Newell and Herbert Simon, became the leaders of the research AI for many decades. They and their students wrote programs that, for most people, simply amazing: computers are solving problems of words in algebra, proving logical theorems and speaking English. In the mid 1960s, research in the U.S. was heavily funded by the Defense Department and the laboratories had been established worldwide. AI Founders were deeply optimistic about the future of the new field: Herbert Simon predicted that "the machines will be capable, within twenty years, of doing any work a man can do" and Marvin Minsky agreed, writing that " within a generation … the problem of creating 'artificial intelligence' will substantially solved. "
They did not recognize the difficulty of some of the problems they face. In 1974, responding to criticism of England's Sir James Lighthill and ongoing pressure from Congress to fund more productive projects, the U.S. and British governments cut off all undirected, exploratory research in AI. In coming years, when funding project was difficult to find, later to be called an "AI winter".
In the 1980s, AI research was revived by the commercial success of expert systems, a form of AI program that simulates the knowledge and skills to analyze one or more human experts. In 1985 the market AI had reached over one billion dollars. Meanwhile, Japan's fifth computer project inspired a generation of U.S. and British governments to restore funding for academic research in the field. However, from the collapse of the Lisp machine market in 1987, AI once again fell into disrepute, and a second, longer duration of AI winter began.
In the 1990s and the 21st century AI achieved its greatest success, although a little behind the scenes. The artificial intelligence is used for logistics, data mining, medical diagnosis and many other areas across the industry technology. The success was due to several factors: the incredible power of computers today (see Moore's Law), a greater emphasis on solving subproblems Specifically, the creation of new ties between AI and other fields working on similar problems, and above all a new commitment by researchers to mathematical methods sound and rigorous scientific standards.
Problems
The problem of simulation (or create) the intelligence was split into a series specific subgroups of problems. These consist of particular features or capabilities that researchers would like an intelligent system to display. Features described below have received the most attention.
Deduction, reasoning, problem solving
Early investigators of AI developed algorithms that imitated the step by step the reasoning that humans use when they solve puzzles, board games or make logical deductions. In late 1980 and 90, the AI research has also developed highly successful methods for treating uncertain or incomplete information, employing concepts of probability and economics.
For difficult problems, most of these algorithms may require huge computational resources – most experience a "combinatorial explosion": the amount of memory or computer time required becomes astronomical when the problem goes beyond a certain size. The search for the solution of problems more efficient algorithms is a high priority for AI research.
Human beings solve most problems with the quick use, trials intuitive rather than conscious step by step deduction that early AI research was able to model. AI has made some progress in the imitation of such "sub-symbolic" Troubleshooting: embedded approaches emphasize the importance of superior reasoning skills consensus; neural net research attempts to simulate the internal structures of human and animal brain that leads to this ability.
Knowledge Representation
Knowledge Representation and engineering knowledge are fundamental to AI research. Many of the machines are expected to solve problems will require extensive knowledge of the world. Among the things that the AI should represent are: objects, properties, categories and relationships between objects, situations, events, states and time, causes and effects, knowledge about knowledge (what we know about what other people know), and many other domains under investigation. A complete representation of "what that there is an ontology (to borrow a word from the traditional philosophy), of which more than general ontologies are called.
Among the most difficult problems in knowledge representation are:
Default reasoning and the qualification problemMany things that people know take the form of "working hypothesis". For example, if a bird comes into the conversation, people often picture an animal that is the size of a fist, singing and flies. None of these things are true about all birds. John McCarthy identified this problem in 1969 as the qualification problem: for any rule of common sense AI researchers care to represent, tends to be a large number of exceptions. Almost nothing is simply true or false in the way that the abstract logic requires. AI research has explored a range of solutions to this problem. The extent of common sense knowledgeThe number of atomic facts that the average person knows is astronomical. Research projects attempt to build a knowledge base full of common sense knowledge (eg, Cyc) require huge amounts ontological engineering laborious – to be built, by hand, a complex concept at a time. An important goal is to have sufficient equipment to understand the concepts of power learn by reading sources such as the Internet, and thus be able to add their own ontology. The shape of subsymbolic knowledgeMuch some common sense what we know is not represented as "facts" or "statements" that actually could be said aloud. For example, a chess master will avoid chess position in particular because it "feels very exposed" or an art critic can take a look at a statue and instantly realized that it is a forgery. These are insights or trends that are represented in the brain non-consciously and sub-symbolic. Knowledge of this information, supports and provides a context for symbolic knowledge, aware. As with the related problem of the sub-symbolic reasoning, it is expected that places computer AI or intelligence describes how to represent this kind of knowledge.
Planning
Intelligent agents must be able to set goals and achieve them. You need a way to visualize the future (which must have a representation of the state of the world and be able to make predictions about how their actions will change) and be able to make decisions that maximize utility (or "value") of the available options.
In classical planning problems, the agent can assume it's the only thing acting on the world and can be sure what the consequences of their actions can be. However, if this is not true, you should periodically check the world conforms to their predictions and must change your plan as needed, which requires the agent to reason under uncertainty.
The planning agent uses several cooperation and competition from many players to attain a certain goal. Emergent behavior like this is used by evolutionary algorithms and swarm intelligence.
Learning
Machine learning has been central to AI research from the beginning. Unsupervised learning is the ability to find patterns in an input sequence. Supervised learning includes both numerical classification and regression. The classification is used to determine what something belongs to the category, after seeing a series of examples of things in several categories. Regression takes a set of numerical input / output examples and attempts to find a continuous function Outputs generated inputs. In reinforcement learning the agent is rewarded for good answers and penalized for bad ones. These can be analyzed in terms of theory of the decision, using concepts such as utility. Mathematical analysis of machine learning algorithms and their performance is a branch of computer science theory known as computational learning theory.
Natural language processing
Natural language processing offers machines the ability to read and understand the languages spoken by humans. Many researchers hope that sufficiently powerful system of natural language processing would be able to acquire knowledge on their own, reading the existing text available over the Internet. Some direct applications of natural language processing include recovery information (or text mining) and machine translation.
The movement and handling
ASIMO uses intelligent sensors and algorithms to avoid obstacles and navigate the stairs.
The field of robotics is closely related to avian influenza. Intelligence is necessary for robots to be able to handle tasks such as object manipulation and navigation, with sub-problems of location (knowing where he is), mapping (learning what is around you) and the planning of movement (figure out how to get there).
Perception
The perception of the machine is capacity utilization of inputs from sensors (such as cameras, microphones, sonar and other more exotic) to deduce aspects of the world. Vision computer is the ability to analyze visual input. A few selected subproblems are speech recognition, facial recognition and object recognition.
Social intelligence
Kismet, a robot with rudimentary social skills
Emotion and social skills play two roles in an intelligent agent. First Instead, you should be able to predict the actions of others, by understanding their motives and emotional states. (These are elements of game theory, theory of the decision and the ability to model human emotions and perceptual skills to detect emotions.) Also, for good human-computer interaction, an intelligent machine also needs to show emotions. At least to appear friendly and sensitive to interacting with humans. At best, it should have normal emotions himself.
Creativity
Topio, a robot that can play table tennis, developed by TOSY.
A sub-field of AI addresses creativity both in theory (from a philosophical and psychological perspective) and practice (via specific implementations of the systems that generate products that can be considered creative).
General intelligence
Most researchers hope their work will eventually be incorporated on a machine with general intelligence (known as strong AI), combining all the skills above and exceeding human capabilities in most or all of them. A few believe that anthropomorphic features like artificial consciousness or an artificial brain may be necessary for this project.
Many AI of the above problems are considered-complete: to solve a problem, we must solve them all. For example, even a single, specific task, such as translation machine requires the machine to follow the author's argument (reason), know what you're talking about (knowledge), and the faithful reproduction of the intention the author (social intelligence). Machine translation, therefore, is believed to be AI-complete: it may require strong AI to be done, and humans can do.
Approaches
There is no unifying theory or paradigm that guides AI research. Researchers agree on many issues. Some of the oldest questions remained unanswered are: artificial intelligence to simulate the natural intelligence through the study of psychology or neurology? Or human biology is irrelevant to AI research as the biology of birds is aeronautical engineering? Intelligent behavior can be described using simple and elegant principles (such as logic or optimization)? Or is that necessarily require the solution of a large number of completely unrelated problems? Can be reproduced using high-level intelligence symbols, like words and ideas? Or a need to "sub-symbolic" transformation?
Cybernetics and simulation of the brain
There is no consensus on how closely the brain must be simulated.
In the years 1940 and 1950, a number of researchers explored the relationship between neurology, information theory and cybernetics. Some of them built machines they use electronic networks to display rudimentary intelligence, like turtles W. Walter Gray & the Beast at Johns Hopkins. Many of these researchers met for meetings of the Society teleological Princeton University and the Ratio Club in England. In 1960, this approach was largely abandoned, although the elements that would revived in the 1980s.
Symbolic
When access to digital computers was possible in the mid-1950s, AI research began exploring the possibility that human intelligence can be reduced to the manipulation of symbols. The investigation focused on three institutions: CMU, Stanford and MIT, and each developed his own style of research. John Haugeland called these approaches to IA "good old AI" or "BAIA".
Cognitive simulationEconomist Herbert Simon and Alan Newell studied the ability to solve human problems and attempted to formalize, and his work laid the foundations for the field of intelligence artificial, as well as cognitive science, operations research and management science. His research team conducted psychological experiments to demonstrate the similarities between the solution of human problems and programs (such as their "General Problem Solver") that were developing. This tradition, focusing on Carnegie Mellon University finally culminate in the development of the Soar architecture in the middle 80s. Logic basedUnlike Newell and Simon, John McCarthy felt that the machines was not necessary to simulate human thought, but should try to find the essence of abstract reasoning and problem solving, regardless of whether people use the same algorithms. His lab at Stanford (SAIL) has focused on the use of formal logic to solve a wide variety of problems, including knowledge representation, planning and learning. The logic also focus the work of the University of Edinburgh and other parts of Europe, which led to the development Prolog programming language and the science of logic programming. "Anti-logic" or "messy" Researchers at MIT (as Marvin Minsky and Seymour Papert) found that difficult problems in vision and natural language processing requires ad hoc solutions – which argued there was no simple and general principle (like logic) that captures all aspects of intelligent behavior. Roger Schank describes his "anti-logic" approaches as "messy" (as opposed to paradigms "clean" at CMU and Stanford). Knowledge bases of common sense (as Cyc Doug Lenat) are an example of "scruffy" AI, since they must be built by hand, a complex concept at a time. Computer knowledge with great memories basedWhen became available around 1970, researchers from the three traditions began to build knowledge in AI applications. This "knowledge revolution", directed the development and deployment of expert systems (introduced by Edward Feigenbaum), the first truly successful form of artificial intelligence software. The revolution knowledge was also driven by the belief that vast amounts of knowledge would be required for many simple applications of AI.
Sub-symbolic
During the 1960s, symbolic approaches had achieved great success in simulating high-level thinking in small demonstration programs. The cyber-based approaches or neural networks were abandoned or pushed into the background. In the 1980s, however, advances in symbolic AI seemed and since many believed that the symbolic systems would never be able to mimic all the processes of human cognition, especially the perception, robotics, learning and pattern recognition. A number of researchers began looking at "sub-symbolic" approaches to specific problems of avian influenza.
Of bottom up, embodied, situated, behavior-based AIResearchers nouvelle or related field of robotics, such as Rodney Brooks, rejected the symbolic AI and focused on basic engineering problems that will allow robots to move and survive. His work is not revived the symbolic point of view of researchers in the cybernetics in the early 50s and re-introduced the use of control theory in AI. These approaches are conceptually related to the embodied mind thesis. IntelligenceInterest Computational neural networks and "connectionism" was revived by David Rumelhart and others in the 1980 East. These and other sub-symbolic approaches, such as fuzzy systems and evolutionary computing, are now studied together for the emerging discipline of computational intelligence.
Statistics
In the 1990s, AI researchers developed sophisticated mathematical tools to solve specific subproblems. These tools are truly scientific, in the sense that their results are quantifiable and verifiable, and have been responsible for many of the recent successes of AI. Shared mathematical language has also allowed a high level of collaboration with more established fields (mathematics, economics or research). Stuart Russell and Peter Norvig describe this movement as nothing less than a "revolution" and "the victory of the neats.
Integrating approaches
Paradigm intelligent agent intelligent agent is a system that perceives its environment and takes actions which maximizes its chances of success. Intelligent agents are simple programs that solve specific problems. Intelligent agents are more complicated and rational thinking human beings. The paradigm of the license provides researchers to study the problems and find solutions which are verifiable and useful, without agreeing on a single approach. An agent that solves a specific problem, you can use any method works – some agents are symbolic and logical, some are sub-symbolic neural networks and others can use new approaches. The paradigm also offers researchers a common language to communicate with other fields, such as decision theory and economics, which also use the concepts of abstract agents. The paradigm of agent Smart was widely accepted during the 1990s. ArchitecturesResearchers agent architectures and cognitive systems have been designed to build intelligent systems intelligent agents that interact in a multi-agent system. A system with both symbolic and sub-symbolic component is a hybrid intelligent system, and the study of these systems is artificial intelligence systems integration. A hierarchical control system provides a bridge between the sub-symbolic AI at their lowest levels reactive AI and traditional symbolic highest levels, where time constraints relaxed to allow planning and modeling the world. Architecture Rodney Brooks subsumption 'is a proposal early in the hierarchy.
Tools
During 50 years of research, Amnesty International has developed a number of tools to solve the hardest problems in computer science. Some of the more general of these methods are discussed below.
Search and optimization
Many AI problems can be solved in the theory intelligent search through many possible solutions: reasoning can be reduced to conducting a search. For example, the logical test can be considered as the search for a path that leads from premises to conclusions, where each step is the application of an inference rule. Planning search algorithms through the trees of goals and subgoals, trying to find a path to a target goal, a process called means-ends analysis. Algorithms move Robotics limb and grasp objects using local search in configuration space. Many learning algorithms use search algorithms based on optimization.
Exhaustive searches are rarely simple enough for most real world problems: the search for (the space the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never complete. The solution to many problems is to use "heuristic" or "golden rules" to eliminate the options that are unlikely to lead to the target (called "pruning the tree search). Heuristics provide the program to guess a "better" than the path of the solution comes from.
A very different type of search rose to fame in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to start the search with any response, then gradually improve the guess until no improvements can be made. These algorithms can be viewed as blind people climbing the hill, we began the search at a random point in the landscape, and then by jumps or steps, we are moving our conjecture uphill until you reach the top. Other algorithms, simulated annealing optimization, search road and random optimization.
Evolutionary computation uses a form of search optimization. For example, one can start with a population of organisms (the assumptions) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining assumptions). The forms of evolutionary computation include swarm intelligence algorithms (such as ant colony and particle swarm optimization) and evolutionary algorithms (eg genetic algorithms [103] and genetic programming [104] [105]).
Logic
Logic was introduced in the investigation AI by John McCarthy in his 1958 proposal Taker Council. The logic is used for knowledge representation and problem solving, but can be applied to other problems. For example, the algorithm uses satplan planning logic and inductive logic programming is a method for learning.
Several different forms of logic used in AI research. Propositional or declarative logic is the logic of statements can be true or false. Logic first-order also allows the use of quantifiers and predicates, and can express facts about objects, their properties and their relationships with others. The logic blurred, is a version of first order logic that allows the truth of a statement to be represented as a value between 0 and 1, instead of simply True (1) or False (0). Fuzzy Systems can be used for uncertain reasoning have been widely used in modern industrial and consumer product control. Default logic, non-monotonic logics and circumscription are forms of logic designed to help with default reasoning and the problem of qualification. Several extensions of the logic has been designed to handle specific domains of knowledge such as: description logics, situation calculus, the calculus of events and calculation of fluid (for the representation of events and time), the calculation of causation, calculation of beliefs, and modal logic.
In 1963, J. Alan Robinson discovered a simple, complete and fully algorithmic logical deduction that can be done easily by digital computers. However, a naive implementation of the algorithm quickly leads to a combinatorial explosion or an infinite loop. In 1974, Robert Kowalski suggested to represent the logical expressions, as Horn clauses (statements in the form of rules: "if p then q"), which reduced the deduction logic of backward chaining or forward chaining. This very relieved (but not eliminate) the problem.
Probabilistic reasoning methods uncertain
Many of the problems of AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. From the late 80s and 90s, Judea Pearl and others, defended the use of methods borrowed from probability theory and the economy to develop a set of powerful tools to solve these problems.
Bayesian networks are a general tool that can be used for a large number of problems, reasoning (using the Bayesian inference algorithm), trained (with the expectation maximization algorithm), planning (using decision networks) and perception (using dynamic Bayesian networks). Probabilistic algorithms can also be used for filtering, prediction, smoothing and searching for explanations for the data flows, helping the perception systems for analyzing the processes occurring in time (for example, Hidden Markov models or Kalman filters).
A key concept of the science of economics is the "utility": a measure of how valuable something is an agent intelligent. Precise mathematical tools have been developed to analyze how an agent can make decisions and plan, using decision theory, analysis decision, the theory of information value. These tools include models such as Markov decision processes, dynamic decision networks, theory games and mechanism design.
Classifiers and statistical learning methods
The simplest form of AI applications can be divided into two types: classifiers ( "if shiny then diamond") and controllers ( "if shiny then pick up"). drivers however, also classified the conditions before deducting the shares, and therefore the classification is a central part of many AI systems. Classifiers are functions that the way matching use to determine a closest match. You can adjust as appropriate, make them very attractive for use in AI. These examples are known as the observations or patterns. In supervised learning, each pattern belongs to a predetermined class. A class can be seen as a decision has to be done. All the combined observations with class labels is known as a data set. When a new observation, that observation is classified based on experience prev.
A classifier can be trained in various forms; There are many statistical and machine learning methods. The most commonly used classifiers are neural networks, kernel methods like support vector machine, k-nearest neighbor algorithm, Gaussian mixture model, Bayes classifier naive, and the decision tree. The performance of these classifiers were compared with a range of tasks. Classifier performance depends greatly on the characteristics data to be classified. There is no single classification that works best on all given problems, which is also known as the "no free lunch theorem". The determination of an appropriate classifier for a given problem is more art than science.
Neural networks
A neural network is a group of nodes interconnected, similar to the vast network of neurons in the human brain.
The study of artificial neural networks began in the decade before the research AI field was founded in the work of Walter Pitts and Warren McCullough. Another important early researchers were Frank Rosenblatt, who invented the perceptron and Paul Werbos who developed the backpropagation algorithm.
The main categories of networks are acyclic or feedforward neural networks (when the signal passes in one direction) and recurrent networks (which allow feedback). Among the most popular networks are feedforward perceptrons, multi-layer perceptrons and radial basis networks. Among the recurrent networks, the most famous is the Hopfield network, a network form of attraction, which was first described by John Hopfield in 1982. The neural networks can be applied to intelligent control problem (for robotics) or learning, using techniques such as Hebbian learning and competitive learning.
Jeff Hawkins argues that research on neural networks has stalled because the model has failed the essential properties of the neocortex, and has proposed a model (hierarchical temporal memory) based on neurological research.
Control theory
Control theory, the grandson of cybernetics, has many important applications, especially in robotics.
Languages
AI researchers have developed several specialized languages for AI research such as Lisp and Prolog.
Evaluate progress
How can you determine if an agent is intelligent? In 1950, Alan Turing proposed a general procedure to test the intelligence of an agent is now known as the Turing test. This procedure allows almost all major problems of artificial intelligence to be tested. However, it is a difficult challenge and currently not all agents.
Artificial intelligence can also be evaluated on specific problems Small problems such as chemistry, handwriting recognition and role playing. These tests were called skilled in the Turing test. Problems offer little more achievable goals and there is an increasing number of positive results.
The general classes of test results of AI are:
- Optimus: You can not do better
- Strong super-human: performs better than all human beings
- Super-human: performs better that most humans
- Sub-human: performs worse than most humans
For example, performance is optimal in drafts, performance in chess is super-human and super-human strength, and performance in many daily tasks performed by humans is sub-human.
A very different approach measures intelligence through testing machine was developed from mathematical definitions of intelligence. Examples of such tests start in the nineties the development of intelligence tests using notions of Kolmogorov complexity and data compression. Definitions Similar artificial intelligence are presented by Marcus Hutter in his book Universal Artificial Intelligence (Springer 2005), an idea developed by Legg and Hutter. Two major advantages of mathematical definitions are its applicability to non-human intelligences and its lack of a requirement for human evaluators.
Applications
Artificial intelligence has been used successfully in a wide range of fields including medical diagnosis, stock trading, control robots, law, scientific discoveries, video games, toys, and Web search engines. Often, when it comes to a technique for general use, no longer artificial intelligence is considered, sometimes described as the effect of AI. Also be integrated into artificial life.
Competitions and awards
There is a series of competitions and prizes to encourage research in artificial intelligence. The main areas of promotion are: machine intelligence in general, the behavior conversation, data mining, driverless cars, robots and soccer games.
Platforms
A platform (or "platform computing ") is defined by Wikipedia as" a sort of hardware architecture or software framework (including application frameworks), which allows run software. "Rodney Brooks, as pointed out many years ago, is not only artificial intelligence software, which defines the characteristics of AI of the platform, but rather the platform itself that affects the IA results, ie we have to solve the problems of AI in the real world platforms and not in isolation.
A wide variety of platforms has enabled the various aspects of AI development, ranging from expert systems, although based on PC, but being around a real system robot global diverse platforms, such as the Roomba widely available open interface.
Philosophy
Artificial intelligence, claiming to be able to recreate the capabilities of the human mind is both a challenge and an inspiration for philosophy. Are there limits to how intelligent machines can be? Is there an essential difference between human intelligence and artificial intelligence? Can a machine have a mind and consciousness? Some of the answers most influential to these questions below.
"Turing polite convention" If an intelligent machine that acts as a being human, then it is as intelligent as a human being. Alan Turing's theory that ultimately we can only judge the intelligence of a machine based in behavior. This theory is the basis of the proposal for the Dartmouth Turing test.The "Every aspect of learning or any other characteristic of the intelligence can be described with such precision that a machine can do to simulate that. "This statement was printed in the proposal for the Dartmouth Conference 1956, and represents the position of most AI work of Simon researchers.Newell and physical symbol system hypothesis "a system of physical symbols have the necessary and sufficient for general intelligent action. "Newell and Simon claim that intelligence is composed of operations formal symbols. Hubert Dreyfus argued that, on the contrary, human experience depends on instinct more unconscious than conscious symbol manipulation and have a "feel" of the situation rather than explicit symbolic knowledge. (See Dreyfus's critique of AI.) Theorem of Gödel's Incompleteness An officer of the system (like a computer program) can not prove all true statements. Roger Penrose is one of those who claim that Gödel's theorem limits what machines can do. (See The Emperor's New Mind). Searle strong AI hypothesis "The appropriately programmed computer with entrances and exits to the right, which had a mind in exactly the same sense of human beings have minds. "Searle counters this assertion with the argument Chinese room, which asks us to look inside the computer and try to find where the "mind" can be.The artificial brain argument, the brain can be simulated. Hans Moravec, Ray Kurzweil and others have argued that it is technologically feasible to copy the brain directly in hardware and software, and that this simulation will be essentially identical to the original.
The speculation and fiction
AI is a common theme in both the science fiction and projections about the future of technology and society. The existence of an artificial intelligence that rivals human intelligence raises difficult ethical issues and the potential power of technology inspires both hope and fear.
Mary Shelley's Frankenstein considered a key issue in the ethics of intelligence artificial: if a machine can be created that has intelligence, you also feel? If you can feel, has the same rights as a human being? The idea also appears in modern science fiction: from the movie Artificial Intelligence: AI considers a machine in the form of a small child has been given the ability to feel human emotions, including, unfortunately, the capacity to suffer. This problem, known now as "robot rights", is currently studying, for example, the California Institute for the future, although many critics believe that the debate is premature.
Another issue explored both by science fiction writers and futurists is the impact of artificial intelligence in society. In fiction, AI has appeared fulfilling many roles, among them;
- As a servant (R2D2 from Star Wars)
- As a law enforcement officer (KITT "Knight Rider")
- Already a colleague (Lieutenant Commander Data in Star Trek)
- As a conqueror / Overlord (Matrix)
- As a dictator (With Folded Hands)
- As a murderer (Terminator)
- As a race sentiant Battlestar Galactica)
- As an extension to human abilities (Ghost in the Shell)
- As El Salvador of the human race (R. Daneel Olivaw in the Foundation series).
Academic sources have considered the consequences such as: a lower demand for human labor, improved human capacity or experience, and the need for a redefinition of identity and basic human values.
Several futurists argue that artificial intelligence goes beyond the limits of progress and fundamentally transform humanity. Ray Kurzweil has used Moore's Law (which describes the relentless exponential improvement in digital technology amazing precision) to calculate that desktop computers have the same processing power as human brains by the year 2029, and in 2045 artificial intelligence will reach a point where it can improve itself at a pace that far exceeds anything imaginable in the past, a scenario writer Vernor Vinge on science fiction called the "technological singularity". Edward Fredkin argues that "artificial intelligence is the next stage in evolution," an idea first proposed by "Samuel Butler's Darwin Among the Machines" (1863), and enlarged by George Dyson in his book of the same name in 1998. Several futuristic Science fiction writers have predicted that humans and machines will merge in the future into cyborgs that are more capable and powerful than either. This idea, called transhumanism, which has roots in Aldous Huxley and Robert Ettinger, is now associated with robot designer Hans Moravec, cybernetics Kevin Warwick and inventor Ray Kurzweil. Transhumanism has been illustrated in fiction, for example in the manga Ghost in the Shell and the science fiction series Dune. Pamela McCorduck writes that these scenarios are expressions of human will to the former, as she calls it, "forge the gods."
About the Author
S. Rajkumar belongs to Madurai, Tamil nadu, India. He is a post graduate in Computer Science and Information Technology. Now he is working as a web designer and PHP programmer in AJ Square Inc. Vilacherry, Madurai.
El Chupacabra caught on tape on the Island of Crete Greece in HD