Artificial Intelligence ----------------------- Artificial Intelligence (AI) is generally considered to be a subfield of computer science that is concerned with the study and representation of knowledge and intelligent reasoning. Additionally one of the main goals of AI is to use the fruits of this study to create computer programs that exhibit, in some fashion, intelligent behavior. Needless to say the field of AI is a vast one that encompasses a huge variety of problem domains, methodologies and applications. Some examples of subject matter that currently falls within the rubric of AI include automated planning, robotics, computer vision, natural language understanding, speech and pattern recognition, neural networks, machine learning, rule-based expert systems, knowledge representation, and automatic theorem proving. The nominal birth of AI is considered to have occurred at a conference held at Dartmouth college in the summer of 1956. The conference was organized by Marvin Minsky, who later helped found the AI laboratory at MIT and who is currently at the MIT Media Laboratory and is famous for his work the Society of Minds. Also present was John McCarthy, the later creator of the LISP programming language that is still an important tool for AI research and development. McCarthy himself at that time named the field artificial intelligence for funding purposes. Herbert Simon and Allen Newell, who had already implemented an automatic theorem proving program at the Rand corporation called the Logic Theorist. These four people are considered the grandfathers of AI. McCarthy went on to later found the AI lab at Stanford while Simon and Newell did the same at Carnegie Mellon. At this conference was assembled a number of other fledgeling researchers who had all written programs of various sorts exhibiting intelligent behavior, and it was the efforts of these individuals that defined and gave impetus to the young field of AI. Also, an important early contribution to AI was given by the brilliant British mathematician Alan Turing in his 1950 paper "Computing Machinery and Intelligence." Here Turing stated the now famous Turing test which stands to this day as an objective measure of intelligence and as a benchmark against which to test new AI algorithms. The Turing test declares a machine to be intelligent if it can pass an imitation game in which a human competitor and the machine itself both communicate with a human interrogator in a separate room only through written questions and responses. If the interrogator cannot, through these written questions and answers, tell the human apart from the machine then the machine is for all practical purposes "intelligent." This simple test has played some role in driving the direction of AI research in the ensuing decades. Much of modern AI research has been concerned with formally capturing the process of reasoning humans use to solve problems, and in creating very general architectures or theoretical frameworks to capture such reasoning. One such paradigm for accomplishing this goal is to view all problems as essentially search problems through a state space of possible partial solutions. The AI system moves from some initial state of affairs, or some initial partial solution to a goal, or final complete solution. The process of moving from partial solution to partial solution is driven by a set of rules which can be thought of as a set of if-then statements. If a partial solution satisfies the hypothesis of the rule, or the condition of the if-statement, then that rule triggers, and the conclusion, or "then" part of the rule transforms the given partial solution into one that is closer to the goal. This type of system is called a rule-based or production system (production because each rule produces a different partial solution). This type of reasoning, namely matching a given partial solution to the hypotheses of different rules is known as forward chaining or deduction. Production systems can also support a different type of reasoning known as backward chaining, where the conclusions of various rules are compared to the desired goal in the hopes of transforming the goal backwards into the initial partial solution. A remarkable number of problems can be solved using this rule-based framework, provided that the representation of knowledge in the encoding of partial solutions, and the rules themselves are chosen judiciously. One of the first examples of such a rule-based system was a simple software robot capable of moving around blocks in an imaginary block world. The state of the world was encoded by relations between blocks, namely a list of which blocks were stacked on each other and which block, if any, the robot arm had in its possession. The goal solution would a prespecified configuration of blocks, and rules were given to allow the robot arm to rearrange blocks from any start configuration to achieve the final configuration without toppling any towers of blocks. A more realistic application is the well known system called MYCIN which is capable of diagnosing bacterial infections. The input to MYCIN is the set of symptoms a patient experiences, and based on a set of rules that are triggered by each symptom, MYCIN returns the particular type of bacteria that could be the cause of the problem. Such an that has a limited domain of expertise, in this case bacterial infections, is know AI system n as an expert system. Although expert systems could not pass the Turing test because of their limited domain of expertise, they have nevertheless become an important part of AI research because of their immediate commercial value. Another aspect of AI whose development has received a large impetus from work on rule-based and expert systems is the field of knowledge representation which is concerned with the creation of interesting data structures to efficiently encode the knowledge that is relevant to solving specific problems. A useful set of tools developed in this field is the notion of frames and inheritance, which are data structures and methods useful for organizing specific facts and the relations between them. Whereas rule based systems attempt to formally capture logical reasoning, their underlying architecture bears very little resemblance to how the human mind is implemented as a vast and intricate network of relatively simple processing units called neurons. A whole other aspect of AI, known as parallel distributed algorithms (PDP's), involves the creation of intelligent systems whose architecture mimics the distributed nature of the human brain. Neural networks have been successfully created and trained to accomplish a host of pattern recognition tasks. The mimicing of nature in AI does not stop there. The field of genetic algorithms (GA's) mimics evolution and uses the the process of natural selection to solve complicated optimization and search problems. A GA maintains a whole population of candidate solutions to the optimization problem at any given time. By picking from among the best solutions in the population at any given time according to a prespecified fitness function and by combining or mating these fittest members to produce new and hopefully better offspring, the GA algorithm efficiently explores the search space and yields highly optimized solutions after many iterations or generations. These biologically inspired methods form an interesting complement to the more formal methods that were initially explored near the beginnings of AI. Ultimately AI is an extremely interdisciplinary subject that draws on the methodologies and expertise of many different fields such as computer science, biology, cognitive science, mathematics, linguistics, philosophy and physics. It is likely that knowledge from all of these fields will be useful, even essential, in searching for the holy grail of AI: a machine that can pass the Turing test. But nevertheless, in the last 50 or so years, AI has produced a wealth of intelligent applications that can achieve many tasks from diagnosing disease to beating grandmasters in chess. The insights gained from creating these applications has in turn lead to great advances in the pure understanding of human reasoning and intelligence, and no doubt many future insights as well as commercial applications have yet to come from the promising field of AI.