We hear the term Machine Learning so often. It has been immensely popular in the cyber world today.

Machine learning is the technology which can make computers work without being explicitly programmed. It enables learning ability in information systems/ computers. As per Dr. Yoshua Bengio, the godfather of machine learning, it is that application of artificial intelligence which can make computers acquire knowledge via data, observations as well as worldly interactions, upon which computers will be able to correctly adapt themselves in any new settings.

Let’s look at the comparison between Machine Learning and Traditional Programming:

In Traditional Programming, input data (a limited set of communication by keyboard, mouse, disc, etc..) will be processed by a program which works on a predefined algorithm to produce output/results (can be a sequence of alphanumeric symbols which is the result of processing). Here, decision making is built directly into the program.

In Machine Learning, input data will be of wider range including one dimensional texts, two dimensional objects, three dimensional scenes, sound, touch, etc.; processing includes knowledge representation, search logic, pattern matching, etc.; and output will also be in the form of printed language and synthesized speech, manipulation of physical objects/locomotion i.e. movement in space, etc.. Here, learning ability will be generated on computers through an agent/trigger, in the processing phase.

Every day, numerous Machine Learning algorithms are getting published. But in general, every Machine Learning algorithm consist of the following :

• Representation – The algorithmic form to which the input data will be brought into / represented.
• Evaluation – It is the application of the respective algorithm in which the input data is represented.
• Optimization – Algorithms are always tuned/optimized within a defined parameter. Parameters represent the range of inputs.

In general, we can say that Machine Learning = Representation + Evaluation + Optimization.

Higher automation, Deeper personalization / customization, Cognitive services / Predictive Analytics are going to be part of every aspect of life in the future. Be it marketing, business prediction, risk analytics, technology or medical science, everything and everyone is going to be revolutionized by Machine Learning.