Machine Learning

Machine learning is essentially a sub-field of artificial intelligence. The essence of this subject is to try and develop machines/programs that have the ability to learn and perform actions without explicitly being told to do so (i.e. without being programmed to do so). This is carried out by implementing certain algorithms so that the machine sifts through tons of data to find certain patterns and use them to automatically adjust its own code/algorithm.

The most famous implementation of this field in the past was in the form of the computer, Deep Blue, which was developed by Carnegie Mellon graduates at IBM in order to play chess. Deep Blue got so good at learning patterns that it defeated the then-world champion, Garry Kasparov in a chess match. The way machine learning works is that the programs mainly learn from examples, and as more examples are inputted, the set of rules it abides by, becomes more and more extensive. The general question that this field aims to answer is, “How can we build computer systems that automatically improve with experience, and what are the fundamental laws that govern all learning processes? “

Machine learning can usually be split into three types of learning:

1) Supervised learning :

In which the program is provided with both the questions and the answers related to the specific issue and it has to identify the rule which allows it to get from point A to point B.

2) Unsupervised learning:

The program is not given any answers, rather, it is provided with quite some raw input and is expected to analyse it and come up with patterns of its own. This is more useful to find hidden patterns in data that regular people may not notice.

3)Reinforcement learning:

This is the type of learning where the machine doesn’t have a ‘teacher’ to tell it when it’s right or when to stop, leaving it to figure out on its own what the next logical step would be and to proceed accordingly. Autonomous cars are good examples of this kind of machine learning.

Machine learning is a very versatile area of study and can be interpreted differently according to the underlying field we choose. For example, machine learning from a software engineering viewpoint is useful since it allows us to build programs that essentially work on themselves, relieving us of any more work than is needed. In conclusion, machine learning is a very exciting area of computer science, one which has endless possibilities that can be utilised in a countless number of ways.

Links

1

2

3

Questions:

What’s the most exciting thing related to this field right now?

What are the fields in which machine learning is most useful?

Are there any ethical concerns realted to this field?