Artificial+Intelligence~

[**A**]**__rtificial__** [**I**]**__ntelligence__**

[*]. History of Artificial Intelligence The idea of artificial intelligence(AI) started with initial idea of creating machine intelligence similar to a human. This can be traced back to Egyptian times, but first expanded with the advancement of computers. With the development of storage spaces for computers in 1949, the possibilities were opened up to the development of AI. The first scientist to make the connection between a human cognitive and machine processing was Norbert Wiener. He brought up the “feedback theory”, of how a computer would control the temperature of the house, cooling or warming dependant on the settings. Followed by Newell and Simon in 1955 who created //the Logic Theorist//, which was the first AI system. It was based from a tree model, each problem put to a branch. It would then choose the branch with the most accurate conclusion. This provided a boost to advancing the concept of artificial intelligence. In 1956, the father of AI, John McCarthy, spread the idea of artificial intelligence to others. He started the branch of computer science now regarded as Artificial Intelligence. As AI caught on to other computer scientists and groups, a program called //the General Problem Solver// was developed along side with the Logic Theorist. Instead of using a tree model, the General Problem Solver was based on a Wiener’s principal of feedback. This enabled more complex problems. McCarthy later on passed the milestone of AI, creating a new computer language called LISP, which stands for “LISt Processing”. This language is still used today, and is a popular choice for AI programmers.

In 1963, the Department of Defence Advanced Research Projects Agency (DARPA)granted Massachusetts Institute of Technology funding to research on “Machine-Aided Cognition”, a form of artificial intelligence. This was to ensure the lead of US over the Soviet Union in the technological race.

The next following years resulted in multitude of AI programs. The most notable was SHRDLU, which used a microworlds method. It would solve spatial problems and logic problems by creating a small world to visualize it. Other notable programs include STUDENT, an algebra solving program, and SIR, a English articulating program. Through the devlopment of such programs, AI has progressed to becoming more and more human like.

The development of Expert systems in the 1970’s allowed conditions and rules to be implemented into the machines. This allowed the computers to understand and forecast stock trends, diagnose of diseases and guide mining explorations.

[*]. Branches of Artificial Intelligence There are many sections of study under the general topic of artificial intelligence. These “branches” can be considered as concepts of artificial intelligence research. **( **** ╯ ****° ****□ ****° **** ） **** ╯︵ **** ┻━┻ ** 1. **Ontology:** “A specification of a conceptualization” (Gruber). Ontology is the description of a set of theories or concepts and the study of the relationships of these concepts. In the reference of artificial intelligence, ontology is used to represent knowledge in various ways. It is best to illustrate it with the situation of a glass that is filled half way. The concept in this case is how to perceive the situation: is the glass half-full, or half-empty, or filled with half of its volume? There are many answers and all of them are correct. Thus, all of these facts can be categorized in one set of a concept. With ontology, computers can use this theory to present answers and knowledge in different ways—all of them can be correct.

2. **Epistemology:** This is the study of knowledge that can be useful in the study of problem-solving. It focuses on how computers and people determine whether if a piece of knowledge or fact is accurate or false. There are four main questions in epistemology: what is knowledge? How is knowledge required? What do people know? How do we know what we know? With epistemology, computers may be able to determine from what is right and what is wrong.

3. **Heuristics:** This is a branch of study that focuses on problem-solving based on past experiences. Humans improve and develop because we learn and discover from our past mistakes and experiences. The same concept can be applied in computers; thus heuristics focuses on the experience-based techniques of solving problems. This can improve computer programs by allowing it to create an alternative solution with the same level of efficiency.

4. **Pattern Recognition:** Certain computers with artificial intelligence nowadays can identify objects by observing any patterns the object provides and studying the data it retrieves. With collected information, the computer can compare it with already known knowledge and identify the object. A prime example of pattern recognition is face recognition—where the computer studies the patterns of someone’s face and identify him or her.

5. **Planning:** This focuses on how computers use facts about a certain situation and use them to generate the best plan to achieve a certain goal in that certain situation. This involves the knowledge of effects of actions, which helps the program to decide what to do. It is the opposite concept of heuristics.

6. **Natural Language Processing:** This concept is all about allowing computers to understand and respond to “natural language”, which is daily speech—informal and “human”. There are many ways to communicate one idea or thought and researchers are finding ways that lets computers to understand all possible speech expressions.

[*]. Approaches of Artificial Intelligence There are many methods to achieve “artificial intelligence” and after years of developing and arguing which method is the best way, two general approaches were standardized. These rivalling tactics are called “Top-Down” and “Bottom-Up”. **( **** ╯ ****° ****□ ****° **** ） **** ╯︵ **** ┻━┻ ** // ^ //** Top-Down: ** This approach simulates the human brain by using gathered knowledge, data, rules, logics, statistics and any other information to solve problems. These collected information is harvested by using separate multiple computer programs and combining their functions, strengths and specific abilities to work together and problem-solve. This concept is also called the “Expert System”.



// ^ //** Bottom-Up: ** This approach is based on the concept of the human brain—which is the networking of neuron cells. Researches must understand the complexities of neurons so that they can mimic the functions with electronic circuits. Warren McCulloch, a Yale medical graduate, and Walter Pitts, a mathematician, suggested that the neuron cells could be considered as devices that process binary numbers. In the same notion, computers also process and work with binary numbers. This is the link between the human brain and a computer-generated brain—the link dubbed “Parallel Computing”. The human body knows what to do because the brain sends signals all around through neuron cells. Boole’s Principle shows how basic logical decisions are made by questioning whether if a fact is true or false (binary numbers). Using the mentioned ideas, McCulloch and Pitts suggested that the concept of neuron-networks may be the basis of developing artificial intelligence. Information can be sent with electronic circuits (acting as connecting cells) and if a “neuron” fails to send the impulse, then the computer is allowed to conclude a “false” answer. The problem with this approach is the architectural and technical difficulties in mimicking the human brain—as there are millions of neuron cells and many connections, which prove to be difficult to model with current materials and resources.



// ^ // Generally, the major difference is that one actually imitates the human brain (Bottum-Up) while the other uses individual programs—known knowledge—to solve problems (Top-Down).

[*]. Computers with Artificial Intelligence
__**Deep Blue (IBM)**__ Deep Blue is known as the world’s best chess-playing computer. In 1996, Deep Blue went up against the world’s best chess-playing human, Gary Kasparov, and Deep Blue managed to defeat him which caused an shocking discovery of saying that computer’s now are at the level of outsmarting human intelligence. Each participant have two hours to make the first 40 moves, which was the common rate of play in human competition, and the winner would receive $400,000 out of a $500,000 purse. Experts were astounded when the computer demolished Kasparov and made him resign on the 37th move in the first game. The first step was the development of an algorithm for constructing chess programs and the programmers programmed the computer certain rules for it should follow. Then the programmer enhancement the AI even further by giving it the ability to analyze numerous positions quickly, which allows it to know every different possible moves based on the situation.



__**Watson (IBM)**__ Watson is known as a supercomputer designed by IBM that won against two jeopardy rivals, which caused a shocking discovery to the world by proving how smarter AI’s are than human intelligence. The programmers designed Watson to show the world how a computer system is able to analyze and process natural language, and then able to predict answers or predictions. Just like humans, Watson relies greatly on content to be able to predict and answer. The machine allows the computer to become smarter and learn as it goes as it answers questions and also learns as it gets the answer wrong or right. Since Watson has the capacity for knowledge representation and reasoning and learn by content and employing language flexibility, it can hold tons of information from dictionaries, plays, books, and encyclopedia.

__**Darpa Grand Challenge (DARPA)**__

The Darpa Grand Challenge is known as a contest that is specifically looking for designs that can be used on rough terrain and for industrial disasters. They have to design a robot that has the agility to maneuver itself into the driver seat and operate an open-frame vehicle such as a tractor. The robots will also have to be able to drive that open frame vehicle to a building and then get out of the vehicle. The challenge says the robots would need to be able to unlock a locked door using a key, walk through the open door, and then walk down a 100m long hallway with rubble obstacles. Once at the end of the hallway, the challenge would be to have the robot climb ladder, locate a leaky pipe, and then close that leak by turning a nearby valve. The final stage of the competition would have the robots replacing a pump so the facility could resume normal operations. The race is really difficult competition as it involves an AI that can meet these requirements.

[*]. Bibliography

“Approaches: Methods used to create intelligence.” __Artificial Intelligence__. June 21 1997. Oracle ThinkQuest. Apr 27 2012 .

Arthur, Charles. “Computers just can’t seem to get past Go.” __theGuardian__. Aug 3 2006. Guardian News and Media Limited. Apr 27 2012 .

“Branches of Artificial Intelligence.” __ISCID Encyclopedia of Science and Philosophy__. International Society for Complexity, Information, and Design. Apr 27 2012 .

"Deep Blue: Year In Review 1996." //Britannica Book of the Year, 1997. Encyclopædia Britannica Online//. Encyclopædia Britannica Inc., 2012. Web. 27 Apr. 2012. <[|__http://www.britannica.com/EBchecked/topic/1368744/Deep-Blue-Year-In-Review-1996__]>.

Gruber, Tom. “What is an Ontology?” Stanford University. Apr 27 2012 .

Hamrita, Takoi. “AI Glossary.” __Engineering Artificial Intelligence__. Apr 27 2012 .

Kawamoto, Dawn. "How IBM Built Watson, Its 'Jeopard'y-Playing Supercomputer." //Daily Finance//. IBM, 11/02/2011. Web. 27 Apr 2012. .

McGlaun, Shane. "Latest DARPA Grand Challenge is for humanoid robots." //Slash Gear//. N.p., 06/04/2012. Web. 27 Apr 2012. .

McMarthy, John. “Branches of Ai.” __What Is Ai__. Nov 12 2007. Stanford University Apr 27 2012 .

Nahudan, Eiman.“Artificial Intelligence.” __Scribd__. Apr 27 2012 <http://www.scribd.com/doc/20106959/ARTIFICIAL-INTELLIGENCE>.

“Ontologies.” __Association for the Advancement of Artificial Intelligence__. Oct 16 2011. AAAI. Apr 27 2012 <http://aaai.org/AITopics/Ontologies>.

“Pattern Recognition.” __Association for the Advancement of Artificial Intelligence__. Oct 16 2011. AAAI. Apr 27 2012 <http://aaai.org/AITopics/PatternRecognition>.

“What is Artificial Intelligence.” __Iloveindia.com__. Apr 27 2012 <http://lifestyle.iloveindia.com/lounge/what-is-artificial-intelligence-11147.html>.