A little bit of background about me will help to show the context of my journey. I have an educational background in Computer Science. Bachelors’ degree in Information Technology and Masters’ degree in CS (specializations: Software engineering and computer networking). During studies, I took some general courses about machine learning, but not enough to consider myself I know machine learning.
In August 2017, about a couple of years after graduation from university, I started learning deep learning. After some research on the Internet, I came across a free course called Practical Deel Learning for Coders. The link takes you to part one of the course. This course is updated in 2020 and can be found here. If I remember right, when I took the first part of the 2017 course, I was a bit lost with all the jargon.
After few weeks going through the course, I realized I need more machine learning knowledge to continue the deep learning course. I did some research again and found Machine Learning course by Andrew Ng on Coursera. I started Machine Learning course on January 2018 and finished it by mid-March 2018 according to my records on Coursera website. This was the right choice for me to start from foundations of machine learning and go to deep learning later.
I am not sure what was the issue at the time, maybe because I was working during the day and taking the lessons after work, I did not continue learning deep learning. I even started the Deep Learning Specialization on Coursera. However, I did not continue much and dropped the whole thing. To start over again in 2022.
I started learning deep learning again in January 2022. The fields of machine learning and deep learning went through fast changes and are still evolving at a quick pace. Therefore, I had to research again and find the most up to date courses and tools. Tools, frameworks, and libraries like PyTorch and TensorFlow can come and go and you need to stay on top of the latest, fastest, best tools.
Not surprisingly, I came across Fast.ai and Deeplearning.ai courses again. I read in a few blogs that fast.ai is more practical than deeplearning.ai. This is true, as I remembered from 2017, fast.ai instructors, have a top-down approach to learning. They do not start with details about mathematics and algebra, but let you run your deep learning classifier in the first lesson.
As mentioned above, fast.ai is updated in 2020 and the course is heavily relied on fastai library. So, I jumped in again to finish what I have started almost 4 years ago. Currently, I am in lesson 3 of the course. The course recommends to start a blog about our journey in learning deep learning.
I started to write about my journey to document my experiences, so hopefully it will be useful for someone else going through this road. Also, I want to write, so I can learn better and have some other people’s opinions about how I am doing and what I can improve. This is a good chance for me to commit in public to be more consistent with learning machine learning and deep learning.
This is the beginning. I will be writing more posts as I go along the lessons and share my learning.