Artificial Intelligence

Abstract

Artificial intelligence (AI), born at the beginning of 1950s, is a young and promising academic subject. Because the question Can machines think? is always disputed, we can not precisely define AI. Usually, we think that AI is a branch of computer science. From the history of AI, we can clearly find that AI together with computer grown up. AI can be widely applied in many practical fields form robots for exploring the planets to automated personal assistants for navigating in the world of networked computer systems.

Keyword：Artificial Intelligence


1 Introduction

Artificial intelligence (AI), as a young and promising branch of knowledge, has attracted a great many researchers’ attention and fast developed in the past 50 years. Nowadays, some artificial intelligence systems have been applied in a wide variety of practical fields. For example, there are programs that generate investment strategies by predicting trends in the stock market, diagnose patient illnesses suggesting treatment, and control robots in factories.

The fast development of AI has gone beyond the most people’s anticipation. The success in biology and modeling of mathematics leads to the foundation and application in computational intelligence systems. These computational intelligence systems refer to the fields of man-made nerve network, evolutionary computation, swarm optimization algorithms and man-made life. Therefore, artificial intelligence can be considered as the combination of some subjects, include computer science, philosophy, physiology, sociology, information science and computation mathematics.

This paper is organized as follows. In Section 2, we define the artificial intelligence. In Section 3, we present the brief history of artificial intelligence. In section 4, we give some practical applications of artificial intelligence. Finally, we will summarize our paper.

2 The Definition of AI

It is difficult to precisely define the artificial intelligence. Artificial intelligence, broadly defined, is concerned with intelligent behavior in artifacts. Intelligent behavior, involves perception, reasoning, learning, communicating and acting in complex environments. By the same time, AI can also be defined as the design and analysis of computer programs that behave intelligently. These programs are constructed to perform as a human or an animal whose behavior we consider intelligent. AI researchers have written programs to control nuclear power plants and diagnose problems in complicated electronic devices. However, it has proved to be more difficult to write programs that can reliably recognize faces or clean our house without destroying furniture.

AI has one of its long-term goals at the development of machines that can do any thing as well as humans can, or possibly even better. At present, the goal of AI is to understand this kind of artificial behaviors whether it occurs in machines or in humans or other animals.

In order to measure artificial intelligence, Alan Turing, a pioneer in the field of computer, once proposed an intelligence test for AI. In the Turing test, a human judge is allowed to ask a program through some sort of an interface such as a video terminal. If the program can fool the human into believing that it is another human responding rather than a computer, then the program is judged intelligent. You can imagine variants of the test in which you manipulate a robot’s environment to see how the robot responds and judge the robot as intelligent or not depending on whether the robot responds in accord with how a human might in the same situation. Now, the Turing test is not considered a very useful test of a machine’s intelligence. For example, Joseph Weizenbaum’s ELIZA program uses some very simple tricks but appears to the tolerant user to be able to carry on a rather realistic. In order to search a less biased arbiter for what constitutes intelligence, many researchers have employed engineering criteria to judge their progress in building AI systems. For instance, if you wish to build a robot to deliver mail, then you can judge the robot’s performance based on its skill in maneuvering in an office environment and its accuracy and speed in routing mail to its intended destination.

3 The Brief History of AI

The first step toward artificial intelligence was taken long ago by Aristotle (384-322 B.C) when he set about to explain and codify certain styles of deductive reasoning that he called syllogisms. Later, some scientists and mathematicians started to pursue automated reasoning. Gottfried Leibniz (1646-1716) invented a calculus philosophicus or ratiocinator in which all knowledge, including moral and metaphysical truths, can some day be brought. Of course, it was a dream that could not be fulfilled with the technical apparatus of the time. Until George Boole [Boole 1854] developed the foundations of propositional logic, automated reasoning had a substantial progress.

In 1958, John McCarthy proposed using the predicate calculus as a language for representing and using knowledge in a system, he called the “advice taker”. This system was to be told what it needed to know rather than programmed. Although, some controversy survived among AI researchers, the predicate calculus and several of its variants constitute the foundation for knowledge representation in AI. During the 1960s and early 1970s, much of the early AI work explored a variety of problem representations, search techniques and general heuristics employing them in computer programs that could solve simple puzzles, play games and retrieve information. One of the influential programs was the General Problem Soler, written by Allen Newell, Cliff Shaw and Herbert Simon.

During the late 1970s and early 1980s, more capable programs had been produced, which contained the knowledge required to mimic expert human performance at several tasks, including diagnosis, design and analysis. The program that is credited with first demonstrating the importance of large amounts of domain-specific knowledge is DENDRAL, a system for predicting the structure of organic molecules given their chemical formula and mass spectrogram analyses.

On May 11, 1997, an IBM program named DEEP BLUE beat the reigning world chess champion, Garry Kasparov, by 3.5 to 2.5 in a six game match. It greatly encouraged AI researchers’ confidence. Projecting present trends into the future, I think that the day when machine can do things as well as humans can is not far away.

4 The Application of AI

Artificial intelligence is the basis for a host of practical systems. Some of these systems have already existed, but the others are being designed for the future. The following presents some applications of AI.

Language translation systems There are now AI translators that you can speak to and have them print transcripts of what you say in foreign languages. The most advanced systems can answer questions based on the information in the text and produce useful summaries.

Air traffic control systems The skies over our airports are becoming crowded as air travel becomes more popular. Tracking thousands of flights, personnel, and maintenance schedules is a difficult job for humans. Computers help by scheduling the arrival and departure of flights to maximize passenger safety and minimize delays.

Supervisory systems As large office buildings and shopping malls become more complex, we need systems to control public services. A supervisory AI system controls elevators, powers and climate conditions. It also manages security and safety inspections and directs visitors to their destinations.

Automated personal assistants It is now possible to build AI systems to actively guide you in using computer networks. They can search through web pages and filter your mail so that you read only the most important and interesting items. They can help you find information, buy products and services, and locate people through the electronic networks.

Intelligent highways Traffic congestion is a growing problem on our major highways. However, doubling the size of our road system would be expensive in the countryside and impossible in cities where space is limited. Instead, efforts are now underway to build AI systems that optimize the usage of existing highways by broadcasting traffic warnings, redirecting vehicles. In the future, your car will plan the route to your destination in coordination with the automated highway managers.

Robots for hazardous conditions Toxic waste cleanup, especially in the nuclear industry, will become increasingly important in the coming years. There are also important applications in biohazard handling in hospital; underground mining; and underwater mining, salvage, construction, and agriculture. Robots can also handle no hazardous tasks that are unpleasant or tedious, such as garbage collection or harvesting crops.

5 Summary

Artificial intelligence is a diverse and rapidly evolving field emphasizing a wide range of techniques and applications. Because the question can machines think？is always disputed, artificial intelligence itself is fully filled with challenge and chance. Therefore, I believe that artificial intelligence will develop quickly and produce a great deal of wealth in the 21 century.