"As soon as it works, no one calls it AI anymore" - John McCarthy
Recently I have started to learn a new field of Computer Science, Machine Learning. But there are so many resources on Machine Learning out there it becomes very difficult to filter out what resources to begin with, so in this post I am going to list out ways one can start learning this new exciting field.
In my opinion, there are 2 kinds of people who want to learn machine learning:
1. People who are interested in maths behind it.
2. People who just want to code and have fun with it.
I definitely fall in the second category as I can do and understand a little bit of math but when I read equations in a ML book or in a research paper I seem to get a little dizzy.
For the people in the first category, I will definitely recommend reading books and research papers and try to understand how everything works, especially, this playlist should help you
This post is for the people in the second category, people who don't want to know a lot of math.
Types in Machine Learning
Machine learning as a whole is divided into 3 types: Unsupervised, Supervised and Reinforcement
This video explains all the types really well, well here's the gist of it if you don't want to watch the video.
Any type of Machine Learning needs data to make the machine learn, well you can give just the data and the model will sort and group the similar data together without any external input, this is Unsupervised Learning, where the data is surely provided but the data is not labeled or given any information on how to group or sort the data, it just groups the data based on similar features.
Supervised Learning is the opposite of it, where we do give the data some labels to nudge the model in the right direction, based on these labels the data is grouped and used by the model.
Reinforcement Learning is where you teach and train a particular model to behave a certain way, the video you watch on youtube of people teaching AI how to walk or play a game like chess or tick tack toe are using Reinforcement Learning.
Ways to start learning ML
First and foremost you need to understand what are you going to learn in ML? From the 3 types elaborated above, For example I am interested in more graphical and fun stuff like teaching AIs how to walk and play, so I am going to go and learn Reinforcement Learning deeply.
For you it may be you are interested in knowing how ChatGPT works or want to make a OCR system from scratch.
Knowing why you want to learn anything is a great way to be interested in learning that in the long run.
Nailing the basics
Whatever you do you must nail the basics. In ML it is having a little basic understanding of how statistic and probability works, for that I would recommend you to have a shot at doing High School Statistics by Khan Academy.
Then you must know how to use a library or some ML algorithms, libraries like scikit-learn would be a great start, this crash course should help you get started.
Then dive deep into one of the types of machine learning. I would recommend start with learning supervised learning as that would cover most of the topics that you would use elsewhere.
Conclusion
Machine Learning is this exciting new field that was growing slowly in the background and has grown to the point that it has started to impact everyone's lives and continuous learning is the only way forward.
Top comments (0)