Machine Learning - An Introduction.

What's Machine Learning?

Machine Learning, shortened as ML is a subset of Artificial Intelligence, that involves the development of computer technologies that don't rely on a set of preprogrammed rules to function, but on self learned data. 

Machine Learning process enables computers to identify relationships and organise patterns and use these identified data to make future decisions on its own.

Types of Machine Learning techniques.

• Supervised Learning. It is also known as Supervised Machine Learning to help distinguish it. In supervised learning, the computer is pre-trained with a set of rules and data to help it categorise dataset and analyse future data using that pretraining. 

An example of supervised learning is the spam detection feature of some applications like eMail apps, messaging applications. These softwares are able to a high degree of accuracy mark certain messages as spam using pre-trained programs.

• Unsupervised Learning. An unsupervised machine learning is trained to connect patterns and give accurate output. This is somewhat similar to how we humans learn new things and use the memory gained from previous experiences to give an accurate answer when in need of it.

 It clusters together similar patterns and gives its output.

Examples of scenarios where unsupervised learning is used is search suggestion.

i.e If user 1(odd number) buys product A and B, User 3(odd number) also buys product C and B, and user 5(another odd number) searches for product B, products A and C will tend to be also suggested to user 5 because the algorithm has clustered some relatable patterns that resonates with him from past experiences.

• Reinforcement Learning. This is the last of the three BASIC machine learning types and it deals with learning its surroundings to be able to interact properly and get desired outcomes. 

Also, this pattern of machine learning is quite similar to how we humans learn from our environment by constantly interacting with it. 

We see a pool, we touch it to really know how it feels like, it feels good, we go ahead to enter into it since it's permeable and we are left gasping for oxygen. We have learnt to associate pools with that experience and whenever we see a pool, we already know that until we learn to swim, we should be careful around it.

This is a simple explanation of how computers use reinforcement learning to learn about its surroundings. 

Reinforcement learning has enabled bots to do more of what humans naturally do while constantly learning. 

Pros Of Machine learning.

√ Less or no Human Intervention Needed.

√ Subject to constant improvement. 

Since machine learning algorithms learn by themselves with little intervention from humans, they're subject to improve as they learn more.

√ Can be used widely. 

Different organisations can adopt machine learning into use.

√ Identify Patterns and Use them in the future. 

Like a Social Media algorithm suggesting posts to us from our browsing patterns which it has learnt.

√ Increase Productivity.

With its fast decision making allowing us to focus more on other things. Imagine sorting out spam mails manually, won't it lead to less Productivity?

Cons of Machine Learning.

√Erroneous. 

Since they still depend on data, and can't literally think for themselves, a wrong data can cause errors in results.

√Loss of employment.

With the reduction of human intervention, people are losing their jobs as these jobs can be handled with more efficiency by ML Technologies. 

√Privacy Issues.

Databases can be hacked and even corrupted leading to serious issues.

√Problem of Cost.

For Machine learning algorithms to be properly set up, huge funds have to be invested in it and also in its maintenance, making it limited to be used by small organisations.