Machine Learning Theory In Artificial Intelligence

Machine Learning

Machine learning is one of the part of Computer science that gives computers ability to learn without being explicitly programmed. Which evolve from study of pattern recognition and computational learning theory in artificial intelligence plus It explores the study and construction of algorithms that can learn from and make predictions or decision.

Machine learning is basically Artificial Intelligence. Rather then making program complicated by entering every data available. We create program that can learn patterns itself. To think like human it needs learning capabilities however it is more than just about learning. It is also about knowledge representation, reasoning, even think that abstract thinking. Machine learning

To pass the test, computer must be able to fool human into believing it is also human. Then in 1952, Arthur Samuel wrote first computer learning program, which was the game of checkers. From that day on it grasp attention of people. In 1990, Machine Learning recognized as separate field and started to flourish. It changed the goal from achieving AI to solving problem of practical nature.

Machine learning can be further defined in 3 categories.

• Supervised Learning - In this system is presented with different example of input and desired output and the goal is to learn from that. So if more examples are given the t will learn more from the data.

• Unsupervised learning – No labels and examples are given to learn algorithm. Assuming it will find pattern or structure from the data. It is long process. It is like teaching newborn baby how to speak. The program does not have any example or set of rules to figure out so it will start from basic. Like it will try to find similarities in data and create structure of data. It is just like learning new language, which we don’t know plus we don’t have dictionary or any other material to analyze, but after some time our mind start understanding pattern of

Higher-level features are derived from lower level features to from a hierarchical representation. It is part of broader family machine learning methods based on learning representation of data. it can observe (Image) will be represented in many ways such as vector of intensity values per pixel, or abstract way as set of edges, region of particular shape, lines and pattern.

Deep learning works in steps like chain system. First set of pair share date with net set of pairs then they will information with next pair. Now consider then face recognition system. In first step it will consider face in box shape and then it will share information to next set of data, which is for eyes and nose. The same thing will go on to next set. By every set the face become clear and till last step we can see the clear face. It will help in recognize the face and texture. We never had technology until now. It will take months to do this process with humans. CPU can do this process in days. And same process can be done in hours with GPU. It can even use to identify different