Machine learning and artificial intelligence have revolutionized the world and are no longer the future. From recognizing objects to being chatbots, artificial intelligence can accomplish tasks that are beyond our imagination. This makes it necessary for us to adapt to the change and have a little idea about machine learning. So in this unfathomable sea of machine learning, where to start?
Just like everyone else, I like to take my own time learning something new. And when I started my self-study journey on machine learning, I gathered resources, reviewed some online lectures, and went through some books. So for you, in this blog on machine learning roadmap, I have summarized how you can start on the journey today and what resources you can use. Below are the topics that are covered in this blog:
Table of Contents
[Note that this is my experience and all the points I state are only my opinions].
The Beginner Course
Three years ago, I decided to start my journey in machine learning and started a course on Coursera. A quick Google search on “free machine learning courses” took me to this course on Machine Learning by Andrew Ng, which I noticed had an excellent review.
I took a glance at the structure of the module and decided to just give it a go without giving it further thought. After all, I just wanted to get started with machine learning so why not with this course! Andrew Ng has done a splendid job of explaining the ML concepts in this course. The following are some of the topics I learned in that beginner machine learning course:
- linear regression
- logistic regression
- backpropagation
- a brief introduction to neural networks
- k-means
- anomaly detection
- optimization techniques like mini-batch gradient descent
Overall, the course covers the most basic terms and concepts that help one get started in the field. This course, however, utilized Octave/MATLAB for the practice assignments. While the important thing is that you understand the machine learning concept itself and not the programming language, programming in Octave is immensely simple and you shouldn’t worry if you have no programming experience.
On the other side, if you have some programming experience in Python like me, you can do the assignments in Python as it will be good practice for programming as well as understanding machine learning concepts. I wanted to get my hands on practicing machine learning using Python for its application in my profession so that is something I followed. This made me set a new goal on my machine-learning path: Building machine-learning programs using Python!
Specialization in Deep Learning
After I completed the course above, I was recommended a follow-up course, again by Andrew Ng, called Deep Learning Specialization. Note that I am talking about the time when I was still in university and had very limited time to focus on such extra-curricular activities. So, this “specialization” consists of five different courses and has an estimated completion time of 4 months at 8 hours per week!
My learning pace is adequate so I considered 1.5 times the time mentioned as I would go back and revise the topics that I learned. That resulted in 12 hours per week! I was already overwhelmed by this number, but hey! You gotta do what you gotta do, right? So I went ahead and clicked on “Enroll”.
The course lasted more than a year in my case. But, was it worth it? Definitely. I got to learn about neural networks, frameworks like TensorFlow, convolutional networks, hyperparameters, and much more. Moreover, I also had an assignment on the YOLO (You Only Look Once), the famous machine-learning algorithm for object detection. Besides, there was an assignment where we needed to write the code for backpropagation from scratch so that was pretty challenging yet an amazing experience.
At the end of the specialization, I was overwhelmed with the amount of machine learning concepts I had learned from those courses. It was certainly clear to me what those algorithms did to produce a particular hypothesis from the data. Now that I had some idea of the code and general knowledge of machine learning, I decided to dive into the mathematics needed for machine learning.
The Mathematics for Machine Learning
Having some knowledge of the concepts of machine learning and deep learning, I started with a book called “Mathematics for Machine Learning”. This is my current read right now and my current position on my road to machine learning. I completed the part on linear algebra and all the topics I read gave me the realization of the math part involved in machine learning. There were many “AHA!” moments while reading that part; it’s like visualizing how the gears get rolling and make the entire machine function. Currently, I am in the chapter on Probabilities and Statistics, and unfortunately, it is my weak spot. I am expecting to spend a significant amount of time on this chapter but looking forward to it. But as a good practice and to understand the concept like conditional probability and Bayer Theorem, I applied the concept on my Instagram account to understand the probability of gaining maximum reach on my posts.
My Future Steps
The subsequent milestone of my machine learning path is to read Hands-On Machine Learning with Scikit-learn, Keras, and TensorFlow. I have looked into the demo of the book and seems to have hands-on practice exercises on machine learning. I believe that solving the exercises in that book with guidance will make me confident in solving actual ML problems.
Conclusion
After having the determination to step into machine learning, I started with the beginner’s Machine Learning course by Andrew Ng on Coursera. I further went on to take the Deep Learning Specialization to understand the intermediate topics of machine learning like CNN and Tuning Hyperparameters. I am reading the book on Mathematics for Machine Learning, and the ultimate goal is to finish the book Hands-On Machine Learning.
One important note I would like to mention is that there is no course or book that will contain everything and a detailed explanation of every concept. While I was on a course or reading a book, I was constantly googling some topics to gain a better understanding. For instance, I kept watching videos of 3 Brown 1 Blue on YouTube, while I was studying linear algebra. Currently, I refer to StatQuest with Josh Starmer parallel to the book to clear my concepts on statistics.
So if you are starting your journey on machine learning, I wish you all the very best. If you are already ahead of me, please share your experience in the comments. If you enjoyed this experience of mine that I’ve documented, get in touch on Instagram. Consider joining my newsletter to stay updated on interesting posts on Machine Learning.
Good and motivating.
All the best with your current course and future plans in ML.
I’m currently doing the ML specialization course by Andrew Ng on coursera.
Hi there. I really appreciate your feedback. Thanks for that. Keep supporting and good luck on your journey in ML. Btw, the course by Andrew is just extraordinary, totally worth it! 🙂
Pingback: 5 Things You Need to Know About YOLOv11 - Machine Learning Site