Machine Learning Roadmap 2019

Machine Learning Roadmap 2019

With so many resources available online, people often gets confused as to where to start their journey on Machine Learning. And then there are those who are not from a Mathematics or CS background, they love to stick to creating models in their browsers. But if you want to build a career in Machine Learning, you cannot escape the clutches of Mathematics or programming. What to do then? Simple, follow this Machine Learning roadmap, use it as a guide, or as a stepping stone to build your career in Machine Learning.

You should start by enrolling in a formal Machine Learning course. And the course that I would suggest to everyone is the one by Andrew Ng. It is available on Coursera, you can audit it for free, or you can pay a minimal fee to get a certificate and access additional materials. Why do I focus on the course by Andrew Ng? Well, Andrew Ng is one of the world's best known AI experts. In 2011, he founded the Coursera. Before that he was head of AI Division at Baidu (the Chinese research engine). Just to give you an idea of how bigger impact Andrew has in AI, Baidu lost $1.5 Billion in value due to his resignation.

This course is highly interactive, you should be ready to spend 5–7 hours per week to get the most out of this course. Each lecture consists of multiple videos with average length of 10–15 minutes. Almost every video has a quiz question to help you make sure that you understand the concept covered in the video. The Andrew Ng course covers the following:

  • Supervised Learning
    Linear regression, logistic regression, neural networks, SVMs

  • Unsupervised Learning
    K-means, PCA, Anomaly detection

  • Special Applications/Topics
    Recommender system, large scale machine learning

  • Advice on building a machine learning system
    Bias/variance, regularization, evaluation of learning algorithm, learning curves, error analysis, ceiling analysis

    Now comes the hard part. How much Mathematics do you need to know? To be very honest with you, Andrew Ng tried his level best to make the Mathematics easier to grasp in his course. However, having said that, you need not worry if you cannot understand everything initially. All you need a refresher course in Mathematics to build a strong foundation. Machine Learning mainly depends on Linear Algebra, Calculus, Probability Theory, Statistics, Information Theory. If you need help, I would recommend 3blue1brown's Essence of Linear Algebra YouTube series. Khan Academy also has a number of helpful videos to clear some Mathematical concepts. You can also try Gilbert Strang's MIT Linear Algebra course.

    What software would you need for the course? Surprisingly, Andrew Ng does not use a programming language like Python or R to teach Machine Learning like every other course does. He uses Matlab (paid software), or Octave (free software) to cover all the materials. By now you must have got an idea about how different his course is. However, if you are comfortable in Python, dibgerge was kind enough to rewrite the exercises completely in Python on GitHub. You can also try this book by Aurelien Geron, an ex-Googler, and is considered the one of the best books on the topic.

    In case you find the course by Andrew Ng to be way more overwhelming than you imagined it to be, you can try out the course on Machine Learning by The only downside to this course is that you need to get hold of a GPU, which doesn't comes for free. You can also try Udacity's Introduction to Machine Learning (UD120). The only downside to this course is that it uses Python 2.7 by default, however, flyinactor91 converted all codes to Python 3 on Github. You have a list of all the changes made on the GitHub repo. If you encounter any errors with this change, which I am sure you would, please refer to this article by Ian Dzindo to learn how to fix those errors.

    What next? After you complete the Andrew Ng course, or the other courses mentioned here, you can move to studying Deep Learning. There are two great MOOCs available online that cover the topic very well:

  • Deep Learning Specialization by Andrew Ng
  • Practical Deep Learning for Coders by Jeremy

    You can also try reading this great blog post by Arvind N. You can also join Kaggle, and participate in their data science competitions. Also, keep yourself updated by following Andriy Burkov, Eric Weber, Matthew Mayo on LinkedIn.

    Start off with Machine Learning today. Keep yourself motivated, you would always find help online.
    « PREV
    NEXT »