From basic Data Science, to Machine Learning, and up to Deep Learning — Getting more theoretical knowledge than practice.
This week begun with the goal of finishing the Deep Learning Fundamentals course at CognitiveClass. The week before I went quickly through several other courses predicated on Python. It was quite easy, because I already had (huge) knowledge of Python.
This time I started struggling. I didn’t have the knowledge of all those mathematical stuff I need to understand the content of the course.
Nevertheless I completed the course successfully and started with Deep Learning with TensorFlow. I thought, that now I will understand it more clearly, because there would be source code provided written in Python. Far wrong!
It turned out, that I was more struggling with understanding the source code than the theoretical stuff. But as a result of needing time to understand the theoretical knowledge, I wasn’t able to understand the source code the other way around.
Based on this experience I changed my learning strategy: As long as I kept busy with programming and all technology topics around it, I was the practical guy. I loved to code around and tried it by myself than reading books about it (although I also read books about it). Now I will have to read and learn more about the theory of Deep Learning as well as about the mathematics to be able to implement it in TensorFlow in the end.
I will number it among the #100DaysOfCode-challenge even if it’s not really coding. But I’m gonna apply it in smaller code-snippets to have at least a little bit of coding in this challenge.
There are only a few things, which have to be implemented, before the functionality is complete. After that I’m gonna make the layout a little prettier, even if the calculator will have a simple layout — design is not my cup of tea!
Furthermore I worked on the home automation and control topic. I passed the sensor data in a logical structure and visualized them properly. Additionally the data is evaluated to provide room summaries: Is the room occupied? Does the room has to be ventilated?
Even if it’s interesting to see what is possible with widely simple data, I’m not quite sure were this journey ends for me.
Besides some batch-scripting experiences on OSX, this week didn’t include more things to write about. It wasn’t as full of coding and work on side-projects as previous ones. Let’s see how the upcoming weeks will be — Keep coding!
Source: Deep Learning on Medium