2013.8.31 - 2013.11.9, 10 weeks
A good introduction to probability, basic concepts and theories, common distributions covered.
https://www.coursera.org/course/prob
Started on 2014.10.6, 6 weeks
Basic OpenGL & GLSL, a ray tracer for the final assignment. This course doesn't teach you how to build computer games, but it does show you some fundamentals of graphics.
https://www.edx.org/course/foundations-computer-graphics-uc-berkeleyx-cs184-1x
Started on 2014.1.6, 12 weeks
An intro course on economics, with plenty of math. The contents are the same as the on-campus Ec1011 course at Caltech.
https://www.edx.org/course/principles-economics-calculus-caltechx-ec1101x
2014.3.17 - 2014.6.2, 11 weeks
Build a compiler for COOL - Classroom Object-Oriented Language (with C++ or Java, your choice) throughout the course. A fascinating course indeed, but be prepared for the heavy load, because the programming assignments are very challenging. This course is pretty much the same as the on-campus CS143 course at Stanford.
https://www.coursera.org/course/compilers
2014.10.2 - 2014.12.1, 10 weeks
Different programming paradigms - especially functional programming, SML, Racket, Ruby used. A very well-designed course, and it has the same contents of the real CSE341 on campus. After taking it, you will come to understand a lot of terms which seems mysterious to you now (functional? closures? ...), and you will be a better programmer for sure.
https://www.coursera.org/course/proglang
2014.9.22 - 2014.12.15, 10 weeks
Usual machine learning methods in practice, with well-designed programming assignments to help you get the feelings. It is much easier than the real CS229 Machine Learning course at Stanford.
https://www.coursera.org/course/ml
2014.12.23 - 2015.2.24, 8 weeks
A much deeper machine learning course, which covers a number of classical machine learning models: SVM (with and without kernels, hard and soft margins), decision tree (random forests), adaboost, neural networks. This course has much more math than Andrew Ng's (some say that Andrew's course isn't even enough as a prerequisite...). It also requires you to code up some of the important algorithms by yourself so that you get some handful practice. This course has the same content as the second half of the ML course on campus at NTU. The first half is offered as another course in Coursera - 'Machine Learning Foundations'.
https://www.coursera.org/course/ntumltwo
2015.1.19 - 2015.3.15, 6 weeks
The first half of the Stanford CS161 course. It covers several well-known algorithms like quicksort and Dijkstra's shortest path. This course focuses on proving the correctness and deriving the running time of an algorithm instead of teaching you how to code it up hand by hand.
https://www.coursera.org/course/algo
2015.1.19 - 2015.3.30, 10 weeks
Dive into the EE world. Fundamental concepts and methods such as DFT, DTFT and FFT covered. Touches linear filters, sampling and quantization. Often uses a guitar as an example to demonstrate.