ARTIFICIAL INTELLIGENCE
Artificial intelligence (AI), the ability to perform tasks typically associated with intelligent beings by a digital computer or computer-controlled robot. The concept is also used in the project of designing structures that are endowed with human intellectual mechanisms, such as the capacity to reason, discover meaning, generalize, or learn from past experience. Since the invention of the digital computer in the 1940s, it has been shown that computers can be programmed with great competence to perform very complex tasks such as finding evidence for mathematical theorems or playing chess. Despite continuous improvements in computer processing speed and memory power, there is still no program that can equal human versatility across broader domains or tasks that require a lot of daily knowledge. On the other hand, in performing such particular tasks, certain algorithms have surpassed the output standards of human experts and specialists, such that artificial intelligence in this restricted context is used in applications as diverse as medical diagnosis, computer search engines, and recognition of speech or handwriting.
Intelligence:
Intelligence is ascribed to all but the simplest human behavior, although even the most complex insect behavior is never taken as an indicator of intelligence. What difference does it make? Consider the behavior of Sphex ichneumoneus, the digger wasp. She first deposits it on the threshold when the female wasp returns to her burrow with food, checks for intruders inside her burrow, and only then if the coast is clear, takes her food inside. If the food is shifted a few inches away from the entrance to her burrow when she is inside the true essence of the wasp's instinctual actions is revealed: on emerging, she will replicate the whole process as soon as the food is displaced. Intellect, obviously lacking in the case of Sphex, must involve the ability to adapt to new situations.
Learning:
As applied to artificial intelligence, there are a variety of different learning styles. Learning by trial and error is the easiest. A simple computer program to solve mate-in-one chess problems, for example, would attempt to move at random until a mate is found. The software could then store the solution with the position so that it would remember the solution the next time the machine noticed the same position. It is reasonably easy to enforce this basic memorization of individual items and procedures, known as rote learning, on a computer. The issue of application of what is called generalization is more difficult. Generalization means adapting the previous experience to similar new circumstances.
Reasoning:
The reason is to draw conclusions suitable for the case. Inferences are either categorized as deductive or inductive. "Fred must be in either the museum or the café," is an example of the former. He is not in the café; therefore, he is in the museum," and of the latter, "Previous incidents of this kind were caused by instrument failure; therefore this accident was caused by instrument failure." The most important difference between these forms of reasoning is that the truth of the premises guarantees the truth of the conclusion in the deductive case, while the truth of the conclusion is in the inductive case." Inductive reasoning is popular in research, where data is gathered and tentative models are produced to explain and forecast future behavior before the model is updated by the appearance of anomalous data. In mathematics and logic, deductive reasoning is prevalent, where complex systems of irrefutable theorems are constructed from a limited collection of simple axioms and laws. In programming computers, there has been considerable progress in drawing inferences, particularly deductive inferences. True reasoning, however, requires more than just drawing inferences; it involves drawing inferences related to the specific task or situation's solution. This is one of the toughest challenges facing AI.
Problem Solving:
Problem-solving can be described as a systematic quest across a range of potential actions to achieve some predefined objective or solution, especially in artificial intelligence. The methods of problem-solving are divided into particular purposes and general purposes. For a particular problem, a special-purpose approach is tailored and sometimes takes advantage of very specific characteristics of the situation in which the problem is embedded. In addition, a general-purpose approach is applicable to a wide range of issues. Mean-end analysis, a step-by-step, or gradual, reduction of the gap between the current state and the final target, is one general-purpose approach used in AI.
Perception:
The world is scanned by means of different sensory organs in perception, actual or artificial, and the scene is decomposed into separate objects in different spatial relationships. An analysis is complicated by the fact that depending on the angle from which it is seen the position and intensity of light in the scene, and how much the object compares with the surrounding area, an object can appear different. Optical sensors to detect people, autonomous vehicles to travel on the open road at moderate speeds, and robots to roam around buildings that collect empty soda cans.
Language:
A language is a system of signs, which by convention has significance. Language should not in this sense, be limited to the spoken word. For example, traffic signs form a mini-language, being a matter of convention that in some countries means "hazard ahead." It is distinctive from languages that linguistic units have conventional meaning, and linguistic meaning is very different from what is considered natural meaning, exemplified in statements such as Those clouds mean rain" and The fall in pressure means the valve is malfunctioning." Unlike birdcalls and traffic signs, their product is an important feature of full-fledged human languages. A productive language can articulate an infinite number of words. It is relatively straightforward to write computer programs that seem able to respond to questions and statements fluently in a human language in severely restricted contexts. Although none of these programs actually understand language, in theory, they may achieve the point where their language command is indistinguishable from that of a regular human being.