Got buzzed with the phrases like A.I. , ML ?? Want to start learning it? This blog is perfect for you.
Machine Learning is a vast field with application in each and every possible domain. So, without proper strategy, your interest will die down soon. So, let’s get started.
Understand why you want to learn it.
This is one of the most important points to consider for learning ML. Do you want to get an extra edge in the company-interviews? Do you aspire to be an entrepreneur and apply ML based technologies to boost up your start-up? I wanted to be a researcher and I found ML as the most useful area to pursue my research.
For example, If you’re preparing for the company placements, a basic knowledge of ML with deep understanding in any one of the sections is enough.
Choose a Proper Course
A proper course is much required to start ML. Most of the ML engineers I’ve spoken with, have started their career with Andrew Ng. (description)
A common mistake among students is that they start directly from Neural network by skipping the initial lectures. Do not do that. One must understand what is the need of NN before starting it.
Make a proper Note
Believe me, the old orthodox style of making notes while listening to the lectures still stands a s one of the most effective ways to understand a topic properly. After covering a few lectures, you can always look back and revise the previous topics.
There are hundreds of useful blogs to help you understand the topics better. One such example is Cola’s Blog. It is followed by most of the ML enthusiasts. Thus, for a particular topic, watch the video, make the proper notes, analyse it, and now strengthen your core by verifying it with the blogs.
Read a lot of Research papers
After you finish the basic topics (say, upto NN), you must start reading lots of papers. A research paper will make you aware of the latest developments in this field. There are thousands of papers on various areas of ML, so choose the papers wisely. Go through the number of citations and the quality of the conference where it is published.
Select your domain
You’ve completed the basic topics of ML, you’ve gone through the various blogs and papers. Now, decide which area you want to go. Most of the learners opt for Computer Vision or Natural Language Processing. Speech Processing is also one of the emerging branches.
Other than these three topics, one can also look for ML in finance, healthcare etc.
Do you have the computational power?
Often we come across codes with heavy computational power requirements for Machine Learning. So, it’s essential for you to have a GPU to train your models. I use NVIDIA 1070 TI for ML and generally prefer everyone to have it. IF you don’t have a GPU, don’t worry. You can train your model in Google Colab. It’s free cloud platform which can be used easily to train your models.
Try to code/ simulate standard results
After learning and understanding basic ML, all you need to do is go through the standard papers and try to code it yourself. Get the help from the corresponding github account. If you get stuck don’t panic. Take help from others or from google.