This is the most common visualization that comes to our mind when we say machine learning. A lot of people relate machine learning as something related to sophisticated bulky machines in general. When thinking in this way, it may seem that Machine Learning requires a great deal of knowledge and a lot of mathematical and conceptual background.
Well, if you think like this, I will surely convince you that training a model can be a really fun and enjoyable activity that can be performed right on your workstation.
What is Machine Learning ?
“Machine Learning is a technique in which a machine is made to perform activities/tasks without being explicitly programmed.”
Machine Learning doesn’t require a lot of technical expertise and high understanding of mathematics to get started. All you need is the curiosity and the enthusiasm in the field of Machine Learning to get started.
Why do we need Machine Learning ?
In earlier times, people used to explicitly code out the rules for a particular application. This code or a set of rules was then fed to a computer which used to perform the tasks based on these rules only.
We all know that the real world is very messy and complex to simulate. Hence new rules kept arising each day. Now, the software developers had to keep refactoring the code on a frequent basis and at times things didn’t use to work well.
If the code used to work well under “environment A”, it doesn’t guarantee it would work in other environments. As a conclusion, we can say that maintaining code with the ever changing rules seemed a very difficult and an inaccurate way of performing this task.
That’s where Machine Learning comes into picture. Machine Learning makes use of Statistical Mathematical Models to work on the past data to draw meaningful insights. Machine Learning models can work with tremendous amounts of data which is far beyond human’s capacity to draw insights based on data visualization.
Does Machine Learning make use of complex Mathematics ?
Well it uses simple mathematical concepts like simple multiplications, additions, summations etc. A basic elementary knowledge would suffice.
Although, there exist some complex “Loss functions” which we will be talking about in upcoming part of this article.
Where can you find Machine Learning ?
Well, Machine Learning techniques can be found everywhere. Believe me, you might not realize that but you are using Machine Learning right away.
If you encountered this article, based on suggestion by Medium because you tend to show interest in Machine Learning, is a classic example of Customer Segmentation.
Similar things can be found on our e-mails, we generally find a lot of mails specifically targeted towards our personal interests. You might also observe that certain e-mails are being classified as spam automatically without your intervention.
The covid-19 estimates and the graph plots we see everyday is backed by Machine Learning.
Where can we use Machine Learning ?
We can use Machine Learning techniques to leverage businesses and startups to make them ready for big technological leaps.
All we need is Data. Data is the fuel for Machine Learning, the better quality of Data one has, really good models can be built out of it.
“Garbage In, Garbage Out” is a very famous quote and it is a very important factor in the field of Machine Learning. If your Data is not meaningful or related to the project you are trying to work, then deploying sophisticated methods and optimizations would lead to bizarre results.
A few applications include :
- Sentimental Analysis of comments or feedback of the customer to find out the issues and resolve them at the earliest.
- Chat bot Systems which can be made running 24 x 7 helps to ease the process of clarifying the queries of the customers.
- Ad detection or Ad Blocker extensions that we use on our browsers.
- Movie Recommendation by Video Streaming Companies like Netflix, Amazon etc.
This list is endless and the opportunities are beyond imagination.
I will cover the types of Machine Learning and the workflow pipeline techniques and the various associated myths in the next article.
We will cover Deep Learning and other famous technology stacks too in the upcoming parts.
Check out my video for the same :