This article is to primarily express the beginning of my Journey through the world of ML/Deep Learning
I was at a customer working on a problem they had with an application, Being in consulting and those of you who has worked with SAP will know that one of first things you do is look for NOTES (bug fixes) to resolve the problem. Since I was from SAP and had access to customers messages, I searched through all the messages in the last 6 months from customers to see if I had missed applying a NOTE. As I was driving back home from the customer that day, something dawned on me. Why should I as a human search through all these past messages to try and find a potential NOTE (bug fixes), we have all the data why cannot we propose a solution the moment customer raises a message, we just need to built some intelligence in the system and propose solutions/knowledge based articles. I was certain that I was not the only one to have hit this problem and thought of a potential solution.
I started trawling through articles on the internet and came across terms such as Machine Learning & Deep Learning. As I started to dig deeper what caught my attention was the application of mathematics/statistics to solve these problems using Computer Programs. I was fairly decent at Mathematics (long forgotten now) and this was a way to channel my interest back in the field along with application to computer science. After some searches I came across Andrew Ng’s course on Coursera for Machine Learning (which was taught on Coursera and Octave was the language used). After finishing the course though I understand the concepts of ML/Deep Learning, I wanted a deeper understanding and more practical real life examples of where these are applied. In the meantime, I was part of consulting and was assigned to a challenging project at a customer that took most of time away from ML/Deep Learning. I call this the “forgetting phase” as I was beginning to forget most of what I learnt in the ML/Deep Learning Course. But I think subconsciously, I always wanted to get back to ML/Deep Learning.
I finally decided to come up with a study plan. By this time I also realised that a vast majority were either using Python or R as a programming language. For reasons (to be detailed later) I choose Python as my programming language of choice for learning ML/Deep Learning. I first wanted to get an understanding of the constructs of Python and get myself familiar. Time was of the essence as I was scheduled to visit my parents in India after more than 3 years for a month. I decided to come up with a learning journey, as a start I decided I am going to tackle python language and get a better understanding of ML, I came across Harrison Kinsley — Python Programming Net that had some basics on Machine Learning using SVM and usage of Python Programming — https://pythonprogramming.net/machine-learning-tutorial-python-introduction/ I decided that before my vacation I will finish this course and then once I am back I will tackle Deep Learning Course on Coursera by Andrew Ng (Bottom Up Approach) & Fast.AI (Top Down Approach) course conducted by Jeremy Howard Rachel Thomas. I plan to post my notes from these courses shortly.
The one thing I would like to point out that there is are tons of material on Web on ML/Deep Learning and it can be overwhelming to know where to start. When I read people posting articles on this is a must read article/book/websites/courses to follow I am still confused? Have I chosen the right learning path? Do I fully understand the concepts? Will I succeed? More deeper questions that I have asked myself Why I am learning ML/Deep Learning what is the purpose? Maybe ML/Deep Learning is not for me, I wont succeed being good at it? I believe each one has to come up with a learning path that suits him/her and answers to these questions. I still have a long way to go and of course I am confused even today, I am learning enough? Have I understood the concepts? Do I understand the math behind? How deep should I go into the math? Do I need a PhD to get into the field of ML/Deep Learning? Maybe since I dont have a PhD this field is not for me? I have read a many articles and visit many websites which I will post below hopefully of some help and not to further muddy the waters. As there are conflicting views on each of these questions. I dont believe there is a right or wrong answer each one of us who is interested in Machine Learning/Deep Learning needs to find his/her own learning path and find their space in Machine Learning/Deep Learning
I had the previlige and pleasure of attending a session in Melbourne by JT Kostman, PhD titled "Applied AI — Coming Soon to a Company Near You…. Preparing for the Near Future of Artificial Intelligence" JT Kostman, PhD It was such a pleasure listening to your insightful thoughts. I came across an article by JT Kostman, PhD on linkedin titled "How to Get a Firm Foundation in Data Science" for some reason I strongly resonate with this article that you need to understand the math behind (how deep, I dont know hopefully some day I will).
I plan to post learnings from my journey and GitHub repository soon….
I would also like to acknowledge many others who have enriched me through their articles/linkedin posts a mention of some them here.
Source: Deep Learning on Medium