Artificial Intelligence: History and General Theory - Assignment Example

Artificial Intelligence; History and General Theory Artificial intelligence or AI; "is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable." McCarthy (2007)1. It is widely believed that the term AI was first coined by John McCarthy at the Dartmouth Conference in 1955 in the famous paper entitled A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence by John McCarthy, M. L. Minsky, N. Rochester, and C.E. Shannon, August 31, 1955: "We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it"2. Centuries before McCarthy coined the term AI, early traces of AI emergence can be found during the time of the ancient Greek Gods; the myths of golden robots created by a Greek God; Hephaestus3. The attempt to create AI could also be seen in the form of sacred statues worshipped in Egypt4 and Greece5. In medieval times, existence of AI was supported by claims made by alchemists e.g. Paracelsus created artificial beings6. Other examples of AI can be seen from the realistic clockwork imitations of human beings built by people like Yan Shi7, Hero of Alexandria8, Al-Jazari9 and Wolfgang von Kompelen10. AI was also present in famous modern fiction such as Mary Shelley's classic Frankenstein and the film Artificial Intelligence:A.I. The modern history of AI includes the birth of computers that were intelligent enough to solve word problems in algebra, proving logical theorems and speaking English. The early 60s saw AI gaining popularity as evidenced by the generous fund provided by the US Department of Defense for AI research. The artificial research however met the doldrums in 1974 with funds on exploratory AI being cut off. The AI research was resuscitated in the early 80s with the commercial success of expert systems that applies knowledge and analytical skills of one or more human experts. By 1985 AI market reached the value of more than a billion dollars11. The AI market again fell into despair with the collapse of the Lisp Machine market in 198712. The highlight of AI success was in the 90s and early 21st century where it was widely used in the technology industry, providing the heavy lifting for logistics, data mining, medical diagnostics and many other areas13. The general theory for AI is mainly based on several characteristics found in normal human being such as deduction, reasoning, problem solving, knowledge representation, planning, motion and manipulation, and perception, amongst others. When solving puzzles, playing board games or make logical deductions a normal human being would go through the process of conscious and step-by-step reasoning14. These traits were successfully imitated through algorithms developed by early AI researchers. Nonetheless, AI researchers have yet to unravel the methods of replicating human traits in solving problems using unconscious reasoning. In order for the machines to solve problems, it needs extensive knowledge about the world e.g. objects, properties, categories and relations between objects15; situations, events, states and time16; causes and effect17. Most knowledge is difficult to be represented due to 3 reasons; default reasoning and the qualification problem, McCarthy [1969]18, unconscious knowledge and the breadth of common sense knowledge. In solving planning problems, the intelligent agent must have the ability of setting objective and achieving it19, thus it needs a way to visualize the future; i.e. having a representation of the state of the world and be able to make predictions about how their actions will change it. The agent must also attempt to determine the utility or the 'value' of the choices available to it20. For robots to be able to manipulate objects and to navigate, it requires intelligence21 such as machine's perception22 which is a machine's ability to use input from sensors (for example microphones and cameras) to see the world and computer vision23 which is a machine's ability to analyze visual input. These abilities can be found in the ASIMO robot which uses sensors and intelligent algorithms to navigate stairs and avoid obstacles. When asked if AI is about simulating human intelligence, McCarthy gave the answer that this is sometimes but not always or even usually the case24. McCarthy elaborated that "on the one hand, we can learn something about how to make machines solve problems by observing other people or just by observing our own methods. On the other hand, most work in AI involves studying the problems the world presents to intelligence rather than studying people or animals. AI researchers are free to use methods that are not observed in people or that involve much more computing than people can do"25. Knowledge and information could be programmed into a computer: was one of two schools of thoughts that came to dominate AI research26. The other view was to express intelligence as formal logic, championed by McCarthy. However, in the last decade or so, a radically different approach which uses statistical tools rather than human-like reasoning has contributed to AI's great success. The concept of "search space", innovated by the late Drs. Newell and Simon, is a way of thinking about possible actions and reactions, establishing an objective, considering all of the possible actions that could be taken and then evaluating which actions are most likely at each step27. Theory on why computers cannot think like a human being was explored by Minsky (2005): "Some computer programs are expert at some games. Other programs can recognize some words. Yet other programs are highly competent at solving certain technical problems. However, each of those programs is specialized, and no existing program today shows the common sense or resourcefulness of a typical two-year-old child-and certainly, no program can yet understand a typical sentence from a child's first-grade storybook. Nor can any program today can look around a room and then identify the things that meet its eyes"28. Minsky disagrees with views that there are aspects of the human brains such as consciousness and sensations that cannot be explained. At the end of his paper, Minsky concluded that "we simply don't yet know enough about what is actually happening in our brains. But when you think enough about anything, then you see this is also the case with everything" 29. Sources: 1. Wikipedia. 2 April 2008. http://en.wikipedia.org/wiki/Artificial_intelligence 2. John McCarthy Home Page. 2 April 2008. http://www-formal.stanford.edu/jmc/ 3. Marvin Minsky Home Page. 2 April 2008. http://web.media.mit.edu/minsky/ Read More