Information about artificial intelligence

The science of artificial intelligence (AI) is trying to comprehend intelligence. Intelligence is the ability to understand a given situation. AI has contributed to many products even in this developmental period. By studying AI we may learn more about human intelligence. The artificial entities are worth studying and are beneficial to humankind. AI covers various areas from perception and logical reasoning to playing chess, math, writing poetry and disease diagnosis. Early science fiction foresaw robots as common by the year 2000. The task proved to be harder than previously imagined. With all of our miraculous technological developments during the 20th century, it seems just a matter of time before we have AI intelligent as the human brain.

There are distinctions between human and computer intelligence. The limitations of computer programs depend on their designer's ability. We do not understand entirely how the human brain functions and so cannot duplicate this to computers. When humans perform better than computers, the programmer’s lack of knowledge shows. Possibly, machine intelligence can be beneficially different to humans.

1. Computers with the ability to think as humans;

2. Computers with the ability to act like humans;

3. Computers with the ability to think rationally;

4. Computers with the ability to act rationally.

Disagreement centres on humans and rationality. That is not to say we are irrational, but we do make mistakes. Human based methods centre on experimental science. The rationalist method involves math and engineering. Each side occasionally criticises the work of the other side. Still, each side has achieved significant outcomes.

Alan Turing developed the Turing Test in 1950 to state when a machine is intelligent. Turing stated intelligence as being able to achieve human level reasoning. A human questions a computer that passes a test if they do not know they are questioning a computer. The computer needs to have the following:

· Language understanding to communicate successfully;

· Be able to store information before or during the questioning;

· Ability to reason, to answer questions and reach conclusions;

· The ability to learn and adjust to a new situation.

· Vision to discern objects;

· Robotics for mobility.

We must determine how humans think before we claim a program can think like a human. Cognitive modelling involves computer AI with human psychology to develop accurate models to test the human mind. With a sufficient theory of the mind, it becomes possible to convey this to a computer program.

Greek philosopher Aristotle was the first to identify reasoning. This began the field of logic. By 1965, programs existed, with enough time and memory spent on a logic problem, solve the problem. AI is trying to build on these programs to make intelligent systems. This is called the Laws of Thought approach. This is no easy task. Taking informal knowledge and changing it to formal logic when the knowledge is uncertain is difficult. Solving a theoretical problem differs from solving it practically.

AI involves the study and building of rational systems. Acting rationally is to reason logically to try to reach one’s goals and then form a conclusion. AI rationality has two advantages. One, the method is more general than the laws of thought method. Two, it is more responsive to science than methods based on human behaviour or thought. Rationality involves clear interpretation.

AI researchers see the need for new basic ideas to achieve duplicated human intelligence. Since the 1940’s the idea of a machine learning like a child seemed possible. This will be true one day. Computers need to increase their knowledge of language enough to be able to learn from reading.

Hitech became the first computer program to beat a champion chess player. Chess strategies only require limited intellectual skills. There lays the limit of the computer’s intelligence. Today’s machines need high computations to understand. Increased knowledge of these mechanisms will lead to fewer computations.

There are many branches of AI. Logical AI involves mathematical logical language in a program to gain knowledge and reach goals. AI search programs often study large numbers of probabilities, eg. chess game moves. Ways are constantly being found to do this proficiently. A pattern recognition program studies and compares what it sees with a pattern. AI matches nose and eyes with a face in a police id composite. Knowledge is represented with a representation program. Epistemology is the study of the types of knowledge needed for solving problems. Planning programs begin with facts on the effects of action, on a given situation and a description of the goal. A generated strategy achieves the goal.

Voice-recognition programs are cutting edge. They do not recognise every word at present, but future refinements will lead to improvements. This is a new aid to allow disabled people to use computers. A robotic system can steer a vehicle. Video cameras, sonar and laser range finders on the vehicle gather input. The system uses these combined inputs with experience gained from training to steer the vehicle. AI and robotics in an unmanned space vehicle aid exploration of an airless planet.

Naturally, there is some anxiety toward advances in AI. Technological progress of the last few decades seems daunting. Computers perform many tasks once thought impossible. Likely around 2020 computers will be more intelligent than humans. Will humanity have lost its unique superiority? This may change humanity and nature significantly. Let us assume we continue superiority. A computer can only function from a power source. We have superiority that enables us to switch the computer off if it does not perform.

Success, unfounded optimism and financial cutbacks dot the history of AI. Inventive methods and constant refining are the best outcomes. Progress in the knowledge of the theoretical basis for intelligence goes with advances in real systems.