Towards the end of last year I started taking the Coursera course “Neural Networks and Deep Learning” taught by Andrew Ng. In a nutshell the course was great. I highly recommend it. That said; I ran into an issue with time. Allow me to explain and how I am addressing it. When done reviewing, will take the next course in the series.
Towards the end of the year, specifically during the last two weeks (i.e., 2018) and the first week of the following year (i.e., 2019) it seemed to be lots of distractions that put a toll on my time allotted to study. This past year the main holidays landed in the middle of the weeks. That meant that social engagements hit on weekends and again during the week.
Every day (including weekends) I get up no later than 05:00 AM (I practice Deep Work. If you wish to learn about the topic click here) and spend the first two hours of the day studying. After that I have breakfast and get ready for the day. Due to the holidays I was not able to get my two hours every day. The material was very infesting and very well presented, but after finishing and passing the course, I decided to do a complete review and then move on to the next one.
One of the items that I enjoyed besides the actual classes, were interviews of people that are leaders in the field of Artificial Intelligence (AI) and specifically Machine Learning (ML). By the way, the term AI is being used less and less due to the bad connotation it has developed during the AI winter. If you wish to read more about it you could do it here.
On three separate occasions Andrew invited Geoffrey Hinton, Pieter Abbeel and Ian Goodfellow. I did notice that during the course, Andrew stated some preferences and suggestions that he uses. Most, if not all the people interviewed on occasions presented their own approach which may or may not have been aligned with his suggestions. By presenting such interviews I got the impression that ML is both a science and an art and is still in evolving. I do appreciate when an individual presents his opinions and the opinions of others.
During the interview with Ian Goodfellow, he mentioned a book that he authored with other two professors and has sold over 70,000 copies. At the end of the interview I looked the book up in Amazon and ordered a copy. Shortly after the book arrived (towards the end of the past holiday season), I started reading it. It seems that for my level of understanding and interests in ML, the first two sections are applicable. This morning I finished reading the first section so I decided to start the review of the course. Hopefully all will be done during this week and I will be able to start the next course next Monday.
I will make sure not to be taking a course next year during the holiday season. I just checked the calendar and it seem that the main holidays land in the middle of the week once again. Will try to finish before the holidays, and take a break of about three weeks or so before starting the following class.
If you have comments or questions regarding this post or if you wish me to cover a particular topic, please leave me a note bellow.
Keep on learning;
Please follow me on Twitter: @john_canessa