1. The main task of Week 1 is to get familiar with machine learning introduction
1) Go through Week 1 of Prof. Andrew Ng's ML course at Cousera.
2) Study Ch.1 of "ML and Cyber Security" book.
2. We also like to set up an experiment environment to run the test codes.
1) Prof. Ng uses Matlab or Octave.
2) Our book use Python.
05/25/2020
This Stanford ML class by Prof. Andrew Ng at Coursera is very good and cover all background we need. It is just starting this week. So, it is perfect for us. I strongly suggest you pay for the certificate. This will be one key learning outcome of our summer project. We can cover it with your professional fee. David may took one similar class before. Just use this as a review and get the certificate. It took me about 4 hours to finish everything for the 1st week. I am starting the 2nd week now. It is very good. You can do it at your own pace, or simply following their weekly schedule. It has 11 weeks.
https://www.coursera.org/learn/machine-learning/
Google CoLab is very convenient, but Google Drive is too slow!! It took too long to upload the 75K inmail.* for spam exercise for Book 1 Ch1. So, I have to install Anaconda on my Ubuntu VM to run the exercise. It took less than 30 minute to set up and run a simple Jupyter Notebook with a browser Firefox on Ubuntu. Easy to run. I haven’t tried Ch.2 exercise yet. I will switch to Anaconda and Jupyter.
Our meeting time: at 8:30pm on Wen every week at https://hawaii.zoom.us/j/3078709726?pwd=OWZLcWl5OVZqNzNwai85SHJmNVdGdz09. NSF may assign other students to our project, starting on June 1st. Let consider them later.
05/23/2020
We can use the Coursera Free Machince Learning Online class offered by Stanford University. https://www.coursera.org/learn/machine-learning/
The book code examples are loaded in the following.
I tried Google Colab. It is very nice to run and easy to use. We can map our Google drive to it easily. But the issue is: it is very slow to deal with many small files. For example, it took me more than 15 hours to upload the 75,000 emails "inmail.*" to my Colab Notebook folder. So, I have to switch to my local machine. I used my Ubuntu 18.04 on my VirtualBox, and installed Anaconda. https://docs.anaconda.com/anaconda/install/linux/ It tooks a little while, but rather straightforward. I then used the simplest Junyper notebook app. It run within seconds, compared to Colab often take minutes or 10+ minutes!!!
I upload two zipped files. Please download to your local machine and unzip them. Let read through the first chapter of the first book. We can discuss later.
Please start go through the tutorial of Google Colab. https://colab.research.google.com/notebooks/intro.ipynb#
It is easy to run on other people’s hosts and we do not have to maintain it. It can map with your Google drive. So, it is easy to work with.
I also attach some notes on Ch.1.
Please do not worry about the ML background. It is the time to learn the basic this summer and hopefully, we can use it for our research.