Contemporary Artificial Intelligence and Vision of It's Future - Research Paper Example

 Running Head: Artificial Intelligence Artificial Intelligence Introduction While prediction of situations that can follow is called intelligence, the same means doing right things at the appropriate time. In the current technological era, artificial intelligence would refer to creating computer software that can perform correct tasks at the right time, while predicting the next challenge it may face in discharging its functions. Accordingly, the ability to sort the correct answer from trillions of available choices could be the modern definition of artificial intelligence.(Intelligence creation) Creating machines, which can engage on behavior that is considered intelligent by human beings, is called artificial intelligence, when related to a particular area of computer science. John McCarthy has coined the phrase of Artificial Intelligence, around three decades ago. However, there has not been any clear definition of this concept, till date. Most common definition is the comparison of computing machines with human brain. Nevertheless, to analyze the areas of artificial intelligence, one may require understanding the concept of intelligence. While many psychologists and cognitive theorists have defined intelligence as the tool that helps in identification of the required knowledge at the appropriate time of decision making; artificial intelligence could be the simulation of human intelligence on a machine. Therefore it can be said that artificial intelligence provides for rational planning and execution of right type of task at the right time. Hence, all computational models, which can think and act rationally, can be linked to AI. However, as any task has two major elements, planning and execution, the planning process of any system should be based on sound reasoning. In addition, any system to perform well would require the knowledge acquired from the real world. Hence, all rational actions must be built-up on the basis of such real world knowledge. In addition, any machine or system must have the learning capability to understand the perceptions involved in the required task. This could provide the basis of thinking and acting rationally, from the machine point of view. Discussion This paper will discuss various aspects of artificial intelligence. These include approaches of AI, examples, branches and applications of AI, along with futuristic vision of the artificial intelligence, with things that need to be done in the coming years, as given in the concluding part of this paper. Some approaches of AI for problem solving While problem solving methods used by AI employ a common term called “state”; it represents the solution status of the problem, at a given step, within the problem solving procedure. Hence, it may need application of an operator to a state for going to the next level of solution and generating next state. The process of having new operators at every level, for creating new states would continue until the desired target is accomplished. Such state-space approach to the problem solving has been the principle of many number puzzle games. While the purpose of such game is to reach the final state, which is the goal; the process involves orientation of digits in the cells to match the game maker’s orientation. In any case, the algorithm of solving state-space problems must concentrate on identifying the required set of operators, while using the same at the appropriate states, during the problem solving process. Search algorithms of AI While water jug problem and traveling salesperson problem are well known classical AI puzzles, certain non-classical problems including diagnosis and pattern classification problems involve non-AI algorithm. However, artificial intelligence approach is totally based on intelligent search and matching, which takes the problem from initial to final stage, while reaching the desired goal. Accordingly intelligent search problems would have the following search algorithms: Generate and test approach As the root of the problem is taken to be the its known starting state, the reasoning space is expanded continuously until the final desired terminal state is achieved. In the process, each generated state is compared and matched with the final state until goal is found, when the algorithm terminates. Whenever the choice of taking multiple paths is given, the algorithm takes the shortest rout to achieve this goal, while generating states and testing the generated states at each level. Hill Climbing Approach This approach is used, many times, while proving trigonometric identities. Under this method, the starting code is selected randomly, without knowing the root code or state, while processing to find the final state in a hillock design. Heuristic Approach As heuristic means the thumb rule, this rule is applied in determining the state, which can suit the problem solving approach, from a number of given states. The quality of this state would ultimately result in reaching the goal, in the shortest possible time, while intermediate states can be reduced considerably. However, heuristic functions are selected on the basis of intuition, with the hope of getting the correct state required for finding the final goal. Means and Ends Analysis As this method aims to reduce the gap between current and goal state, the distance between these two states is measured, for the purpose of applying an appropriate operator to the current sate. This results in reducing the distance between ruling state and the goal state. Many mathematical theorems use this method during their proving process.(Artificial Neural Networks) Examples of artificial intelligence Humans have developed various robotic systems, which are intelligent machines that mimic the human body and perform tasks akin to human functions. Some of these computerized intelligent machines are described below: Humanoid Robot This robot has the appearance of a human body, while it allows the interaction, which humans are supposed to perform with surrounding environment or with tools that are meant for human beings. However, humanoids are different from other kinds of robots, as they are autonomous. This means that these robots can adapt to the changes in the environment, while pursuing their goal. As such they have certain independence in performing desired functions without any human guidance, while working in an unstructured environment. Due to this fact, these robots are able to do tasks that require fair amount of reasoning and intelligence, like cleaning floors, waste-water treatment and even space exploration. Robotics research has produced robots that can cope with the environmental changes, on land, air or water. Accordingly, the capabilities of humanoid would include doing fully autonomous tasks, like charging itself, while conducting self-maintenance. In addition, these robots have the capacity to gain and learn from new environments, as they keep on adjusting to the same, without any human assistance. These robots are also capable of interacting with human beings, in an intelligent way. Being complex creatures, they differ from the other robotic machines, as they can perform the critical tasks of sensing, planning, controlling and actuating a certain task. As humanoids are meant to copy some of the physical and mental tasks for human beings, scientists have used them to create such machines which can have the maximum possible human-like intelligence. The applications of such robots can be seen in the field of engineering, medicine, as well as, cognitive and linguistic fields. These robots have been used to develop human cognition, while understanding the computational model of human behavior. In addition, these humanoids are used for human entertainment also, as they can sing, dance, among performing other tasks. Autonomous humanoids can be used for commercial purposes, as they gain information from the surrounding environment, thereby performing the required tasks to achieve the desired goal. Some common sensors used in creating such intelligent machines include, thermal, optical, contact, sound, touch and electro-magnetic sensors. Such sensors are used to sense the environment and for performing the tasks like moving of lawns, at the desired grass growth level and other such functions. Similarly dirt sensors, in a vacuum cleaner can provide information on the presence of dirt and required time for cleaning the same. More sophisticated tasks can be performed by advanced robots, as they are employed in cruise missiles, unmanned Aerial Vehicles, and unmanned space-ships, among others.(Humanoid robot) Android A robot called android resembles human beings, in appearance and behavior. The examples of androids are fully mechanical robots and cyborg, which is both organic and mechanical. A bio-robot is an android which has been artificially constructed to imitate the human appearance and thinking process. Latest advances in nanotechnology have resulted in creating androids and other robots, at the scale of nanometers. The science of nanorobotics is the technology which creates such machines. These machines are useful in the fields of medical investigation and drug delivery.(android) Thoughts that control cursors Recent developments in the field of computer technology have resulted in the development of cursors that are being controlled only through human thoughts. People can gain control over their brain signals, when a feedback of these signals is provided to them. As functional Magnetic Resonance Imaging (MRI) system registers blood flow in active brain regions, the technique can be used to acquire this feed back, in real time, at a faster speed. The operation of computers through thought –controlled cursors would be a boon to human beings, who are not able to perform simple physical tasks, as well as those with brain injuries. The use of MRI technique can help in humans controlling their pain, while they can learn more about the structures of the disease itself. The process involves computer analysis of brain patterns, while thinking of a particular task. With the help of such analysis, the cursor movements could be changed with different brain patterns, arising from thought changes.(Controlling cursors with thoughts) Branches of Artificial intelligence A few branches of artificial intelligence are briefly detailed below: Logical AI comprises of particular mathematical logical language, which represents the knowledge of facts of a specific situation, as provided in a computer program. Search programs relate to the probing of large number of possibilities, while continuously discovering various methods of executing the program in new domains. Pattern recognition is another branch of AI which requires comparison of observations, within a computer program, with the specific pattern. Such knowledge is helpful in formulating a vision program, which requires sketching a face from the known pattern of eyes and nose. Inference is used for most of default reasoning programs, in which the conclusions are inferred by default. However, new methods involving non-monotonic inference have been added to the logical programs, since last 3-4 decades, whenever ordinary logical default reasoning becomes monotonic. While we have seen considerable progress in developing non-monotonic reasoning and theories of action, since last few decades; the branch of AI which deals with common-sense knowledge and reasoning needs much more to be done. For example application like Cyc system contains many common sense facts, but the same is incomplete, as yet. Similarly epistemology relates to the study of knowledge which is required for solving different problems, while ontology refers to the knowledge which exists in AI. In ontological studies, properties of various kinds of objects are observed in different programs. (McCarthy, 2007) Applications of AI Some of the applications of artificial intelligence are listed below: Speech recognition is one such application, which has proved beneficial for segments like aviation, where the keyboard functions are replaced by a system using speech recognition for flight information. However, use of keyboard and mouse still remains a convenient method of using computer systems. The other important application of AI could be use of intellectual mechanisms, called expert systems, which can provide the desired results. For example, MYCIN was developed in 1974, which could predict the medical diagnosis related to blood infections, while suggesting the required treatments. Playing intelligent games like chess is now possible, as one can buy machines that play master level chess, for a few hundred dollars. The artificial intelligence involved in such games employs brute-force competition principle, while choosing form millions of possible positions.(Applications of AI) “Explicit-implicit interaction” –psychological theory According to Ron Sun, Professor of Cognitive Sciences at Rensselaer Polytechnic Institute, this theory provides an explanation of human mind and its problem solving capabilities in a creative manner. The studies conducted by Sun included the development of a program called, “CLARION”, which acts as a cognitive system. Human beings were introduced to various tests, including coin question, which received responses from human subjects on their thought process. The results showed that 35% of the respondents answered correctly, which indicated the matching of simulation data with the human brain data. Building on the process of “stage decomposition”, the explicit-implicit interaction has been developed in a careful manner, which deals with the creative problem solving aspect of human mind. Stage decomposition process of creative problem solving requires human mind going through four stages for the purpose of finding the correct solutions. These stages are preparation, incubation, insight and verification, as developed by Graham Wallas, in his book, “The Art of Thought” in 1926. Based on Wallas’ works, many theories have come up, which try to explain the specific processes adopted by human mind during the stages of incubation and insight. These theories have been represented in various artificial intelligence models, mathematically. However, it remains a fact that the theory of explicit-implicit interaction is based on the simultaneous existence of two types of knowledge and processing- explicit and implicit. While explicit knowledge could be used for the purpose of reasoning, thinking about different options; the implicit knowledge relates to the intuition that can result in finding the correct solution. Accordingly, both types of knowledge compliment each other and are applied for finding the correct solution. Hence they are important while tackling issues involving creative problem solving techniques.(New Psychology Theory, ScienceDaily, 2010) Conclusion The use of machines, particularly computers, to create artificial intelligence has advanced at a faster sped during the last couple of decades. Applications of robotic engineering has helped in creating intelligence that can prove useful in medical, engineering and mining sectors, which require unmanned investigation at certain times. The benefits of such artificial intelligence is visible as a patient, sitting in a remote corner of the world, is treated via computers, with directions from best physicians across the world. However, there is still more that needs to be done; to ensure that a machine can work as intelligently as human brain, which would make it impossible to distinguish the two, while performing certain task. The explicit-implicit interaction theory mentioned above, briefly, throws some light on the future of artificial intelligence. The knowledge processing, particularly reasoning and thinking requires in-depth study, to ensure that it matches the capabilities of human brain in identifying, selecting and executing an option, from the millions of available choices, at a given time. The Robotic science, being the most advanced stage of artificial intelligence, still needs human brains to analyze the results collected by a robot, from various experiments, like space exploration etc.. BIBLIOGRAPHY WORKS-CITED and sources used Android, Retrieved on Dec 4 2010, from: http://www.sciencedaily.com/articles/a/android.htm Artificial Neural Networks, “Digital Signal Processing-Algorithms and applications”, Artificial intelligence, Retrieved on Dec 4 2010, from: http://www.learnartificialneuralnetworks.com/ai.html “Controlling Cursors with Thoughts: Faster, Simpler and More Accurately”, Nov 19 2010, Retrieved on Dec 4 2010, from: http://www.sciencedaily.com/releases/2010/11/101116102432.htm Humanoid robot, Retrieved on Dec 4 2010, from: http://www.sciencedaily.com/articles/h/humanoid_robot.htm Intelligence creation, “Intelligence is prediction” Retrieved on Dec 4 2010, from: http://www.existor.com/intelligence McCarthy John, Nov 2007, “What is Artificial Intelligence-Branches of AI”, Retrieved on Dec 4 2010, from: http://www-formal.stanford.edu/jmc/whatisai/node2.html McCarthy John, Nov 2007, “What is Artificial Intelligence-Applications of AI”, Retrieved on Dec 4 2010, from: http://www-formal.stanford.edu/jmc/whatisai/node3.html “New psychology theory enables computers to mimic human creativity”, ScienceDaily, Retrieved on Dec 4 2010, from: http://www.sciencedaily.com/releases/2010/12/101201124345.htm Read More