The world’s leading publication for data science, AI, and ML professionals.

Ways you can teach yourself Machine Learning for free at home!!

A list of resources on the Internet.

Photo by Bit Cloud on Unsplash
Photo by Bit Cloud on Unsplash

I started my Machine Learning Journey almost a year ago immediately after I finished reading Pedro Domingos’ book: The Master Algorithm. The book profoundly motivated me to become a part of the current Machine Learning/Deep Learning revolution. So I decided to dedicate a few hours of my daily schedule to learning machine learning. Almost a year in, I have learnt a lot and I am highly obliged to the generous resources on the Internet for making it possible.

Today, I would like to share these free resources, and ways you can use them in order to leap into your Machine Learning Journey. All you need is an Internet connection, basic programming skills (Python), a computer and an adequate amount of motivation. Here we go… note: ‘ * ‘ means recommended to complete

  1. *[fast.ai:](https://www.fast.ai/2020/08/21/fastai2-launch/)**

Fast.ai is a project by Jeremy Howard and his team that aims to make Deep Learning accessible to everyone. This course is quite popular among beginners. Jeremy Howard’s approach to teaching Machine Learning is both intuitive and applicative. On my very first class, I learnt how to build an Image Classifier to classify dog breeds. That motivation alone was enough to keep me going until the end. The fast.ai library is powerful and supported by Kaggle, Google Cloud Platform, Amazon Web Services, Paperspace and several other platforms. They recently launched a fast.ai version 2 which is more flexible, faster and beginner-friendly. Be sure to check out their Deep Learning course. It’s amazing and absolutely free.

  1. Lectures and Video Tutorials on YouTube:

There is an abundance of Machine Learning Courses on YouTube. Stanford, MIT, Caltech, Paul G. Allen School,…all offer outstanding courses. You can even access the lecture slides, **assignments***(good practice) and solutions from the course website. You can access such world-class education from your room. Moreover, you can find courses on foundational subjects like calculus, statistics, probability, linear algebra and advanced topics like computer vision, reinforcement learning, speech and Natural Language Processing.

You can always look for tutorials to get a quick insight or recap of a certain topic. Here is a list of some of these sources:

a. Stanford’s Machine Learning Course by Andrew Yang (Newer Version)* b. Paul G. Allen School’s Machine Learning Course c. MIT Open Courseware-Linear algebra, probability and statistics, calculus and more d. StatQuest with Josh Starmer– great tutorials on Machine Learning e. Two Minute Papers-for quick insights f. CodeEmporium– for quick insights

I would also recommend listening to podcasts by Lex Fridman(on YouTube and Spotify) and Towards Data Science(on Spotify) where they interview experts on Machine Learning and Artificial Intelligence in general. They are insightful and motivational.

4. Blogs, Articles

You can find brilliant articles on almost all machine learning topics online. You can find several articles on implementing models, feature engineering, mathematical intuition, problem-solving, ML libraries, and much more. The following is a list of some very useful websites:

3.Kaggle

Kaggle is a great platform where you can practise your machine learning skills. There are thousands of datasets which you can download and experiment with. Kaggle hosts competitions where you can test your machine learning skills to solve real ML problems. I *highly suggest going through the notebooks and discussions posted by fellow Kagglers. You do not need to be able to understand every detail in the notebooks. You will always learn something new from them no matter what stage you are currently in.**

4.Free Compute

Machine Learning is a computationally resource-intensive process. Buying a decent machine can cost you thousands of dollars. Not every beginner can afford, or is willing to make, such an investment. Fortunately, you can get free cloud computing from a number of online sources: →Kaggle Kaggle notebooks have sufficient computational resources for most machine learning tasks. Most importantly, Kaggle offers its users up to 42 hours of GPU time and 2 hours of TPU time every week. These are decent Tesla K80 GPUs that can speed up most of your deep learning tasks. Unlike the other options on this list, you do not need to give away your credit card information. All you need is a phone number for verification. You can load datasets on Kaggle almost instantly to your notebook folder and most machine learning libraries like NumPy, Pandas, PyTorch, TensorFlow come pre-installed.

Google Cloud Platform (GCP) Free Credits- One year time limit GCP offers $300 worth of free credits that you can use up to 1 year from the date you sign up. You can create and use a decent machine at a $1/hour rate. Combined with Kaggle notebooks, these credits should be enough for a year. I generally used Kaggle notebooks whenever possible and turned to GCP when the compute on Kaggle was not enough.

5.Papers on Machine Learning

Once you get a proper grasp of the concepts of Machine Learning/Deep Learning, you can gradually start reading research papers. You can find most research papers on Machine Learning on arXiv.org. If you are unable to grasp any concept, you will surely find a simplified explanation on some of the blog sites listed above. Try implementing the ideas from the paper to your own projects and observe how they perform.

Considering the amount of information there is on the Internet, this article is by no means complete. The variety the Internet provides can be overwhelming to beginners. Through this article, I wanted to constrain that variety by listing out my personal choices. I hope this is helpful.

Good luck on your journey, it is tough but gratifying😀 !!!


Related Articles