Machine Discovery  (Fall 2016, 3 credits)


Instructor: Prof. Shou-de Lin (

Classroom: CSIE 104

Meeting Time: Mon 2:20-5:20 pm

Office Hour:  After class or by appointment

TA : 楊鈞百,李廣和,許晉齊

Course Description:


It is widely accepted that machines can learn as good as human beings in many tasks: feeding a learning algorithm with sufficient amount of training data, it can generate a model that maps the inputs to the most plausible outputs. Discovery, on the other hand, is about finding something for the first time, which has never been observed in the training data. One interesting question to ask is whether machines can perform discovery when there is no labeled data available. This course will introduce techniques that allow machines to perform discovery. It will also cover several real findings in this topic. Such findings show that with the availability of large unlabeled data and powerful computation infrastructure, machines can indeed perform discovery tasks.




Three Homework assignments (70%)

Final Project (30%)


 Syllabus (Tentativee):

 9月12日 Intro to MD
9月19日 PGM HW1 Out
9月26日 PGM
10月3日 PGM
10月10日 No Class
10月17日 PGM-based Discovery
10月24日 Optimization  HW1 Due
10月31日 Optimization
11月7日 Optimization
11月14日 Optimization-based Discovery
11月21日 No Class HW2 Due, HW3 out
11月28日 Other Learning Models
12月5日 Other Learning Models Based Discovery
12月12日 Final Proposal HW3 Due
12月19日 Knowledge Discovery and Pattern Mining
12月26日 Knowledge Discovery and Pattern Mining
1月2日 No Class
1月9日 Final Presentation