Artificial Intelligence (AI) is a branch of computer science, which tries to give “intelligence to machines“. But the concept of intelligence itself is debatable, and making these machines without life “intelligent” is something near to impossible. But we can say safely that AI aims at producing intelligent behavior from machines. What is the difference between having intelligence and having intelligent behavior? You can exhibit intelligent behavior in a narrow field for some time without really being intelligent. For example a computer playing chess at the master level does not even know that it is playing chess. But for the outsider the view is that it is intelligent like a master. Also we need only this intelligent behavior for many practical purposes.
AI uses ideas from diverse branches of knowledge like computer science, economics, biology, social sciences, mathematics and even grammar. It has also diverse applications in many areas of life. That is, it is an interdisciplinary subject, which takes ideas from almost all fields of knowledge and has applications in many diverse areas of life. Some of the branches of AI are discussed below. This in no way is a complete list.
1. Game playing:
Playing a game like checkers, chess, or go, requires a lot of intelligence for a human being and so these tasks were one of the earliest attractions of AI. Samuel wrote a program playing checkers in the 60’s and many people contributed to the theory of game playing. Finally when a computer could beat the then world champion in a chess game, that was considered a victory of machine over man even though it was not so. Currently there are well-known algorithms for game playing and game playing is considered falling into the domain of algorithms than AI.
2. Automatic theorem proving:
Mathematicians are believed to be super intelligent creatures, and so in its early childhood AI tried to show intelligence by creating machines capable of proving theorems by themselves. By having some basic assumptions and rules, they tried to prove theorems by combining these rules, getting new assumptions and so on. Gelernters program for geometry theorem proving was a typical example. Later it was realized that the intelligence of human experts in this area is not easily imitable, and the use of common sense and knowledge about mathematics was needed to prove theorems. Currently not many developments are taking place in this area.
3.Natural language processing (NLP)
Languages like English, French or Malayalam, which are used by men, are called natural languages. The language we speak is often context sensitive. “Did you shoot the tiger?” has different meanings when asked to a hunter and a photographer. Also our language is incomplete. Natural language processing deals with understanding this language using knowledge about the grammar rules and context. This has a wide variety of applications and is a field of active research. Also translation between these languages is studied in AI.
4. Vision, Speech recognition and similar areas.
Seeing an animal and recognizing it as a cat is child’s play, but a difficult task for computers. Modern AI programs focuses on recognizing objects and persons and behavior based on vision. This has many applications in robot navigation, crime detection, military operations and so on.
5. Expert Systems
Human experts are rare, costly and perishing. If we spend a large sum and train a person as a neurologist, the maximum we can expect is 30-40 years of service. And we cannot take a copy of the neurologist! So if we can train a computer to have the same expertise or to be precise expert behaviour, at least in a narrow field, the utility is high. Expert systems deal with extracting expertise and porting it to computers. That is creating software that can exhibit expert behavior. This field has undergone explosive growth in the last few years.
6.Neural networks
Animal brain is composed of neurons and performs computations (thinking) by passing signals between these networks of neurons. Why not imitate this and evolve intelligence? Neural networks began from this foundation. They are capable of learning, adapting and predicting. Putting in simple language, a neural network is a collection of computation units (real or virtually created), which are interconnected and cooperates for computation. Neural networks have applications in control systems, speech and natural language processing, vision and many other fields.
There are various other areas in AI. It is a vast and emerging field. I will tell more about it in my next article.
Kannan Balakrishnan is a budding indian writer. He continuously writes on a variety of topics like website design, Computer science, self improvement etc. Now his thoughts are presented through celebrated blog http://www.kbwrites.blogspot.com.