The first word that came to my mind when I heard the phrase “Machine Learning” was robots. Indeed, I have had this perception for a long time that machine learning relates solely to robotics, but during the course of my research I’ve come to realize that this is not at all true.
Machine learning is in fact “programming computers to optimize a performance criterion using example data or past experience”, as Ethem Alpaydin, Professor in the Department of Computer Engineering at Bogaziçi University, Istanbul, rightly put forward. In other words, it means that machine learning is a science that explores the study and construction of algorithms that learn from data or some well-defined parameters and thus can make predictions from other data. Machine learning makes use of the theory of statistics to build mathematical models as the core task is to infer from a sample.
I also learnt that the application of machine learning methods to large databases is called data mining. I believe it is called like this due to the analogy made here to mining, that is, the extraction of a large amount of raw materials from a mine, which when processed leads to a small mass if a precious and desired material. Similarly, in data mining, a considerable volume of data is processed to construct a simple model with a valuable use. There are numerous applications of this for instance, analysis of past data in a bank to build models to use in credit applications, fraud detection and the stock market. In the manufacturing industry too, learning models are used for optimization, control and troubleshooting. Moreover, in medicine, learning programs are used for diagnostic purposes, while in telecommunications, call patterns are analyzed for network optimization and to maximize the quality of service. Even in science, the large amount of data in physics, biology and astronomy can only be analyzed at a reasonably fast pace by computers.
During the course of my research, I also realized that machine learning is not just a database problem; in fact, it also forms part of artificial intelligence. For a system which is in a changing environment to be intelligent, it should be able to learn. If the system has the ability to learn and adapt to such changes, then, there is no need for the system designer to foresee and give solutions for all possible solutions.
An interesting thing to note is that machine learning also helps us in finding solutions to many problems in vision, speech recognition, and robotics.
Here are some links which you can refer to to get more information about Machine Learning:
I’ve read somewhere that the learning model can also be “descriptive”. How can it be so? I am not able to understand this.
Can you please explain to us how exactly speech recognition work?
Can you please explain in detail how fraudulent acts can be detected by machines?
Photo Credit: isl.ce.sharif.edu