Semester.ly

Johns Hopkins University | EN.600.676

Machine Learning: Data to Models

3.0

credits

Average Course Rating

(3.88)

Students in the class will be asked to do assignments in Matlab. Matlab is typically easy to pick up if one is already familiar with a different programming language. Students are expected to be mathematically mature. One should have taken at least an introductory course in probability theory and linear algebra. Though not required, exposure to optimization or machine learning is recommended. Proficiency in at least one programming language is expected. When in doubt, send the instructor a copy of your transcript to see if the class is appropriate for you. Also, sit through the first few sessions and first homework to get a sense of fit. Requistes include Intro Prob/Stat, Linear Algebra and Intro Machine Learning as well as strong background in s.

Fall 2013

(3.71)

Spring 2013

(4.24)

Spring 2015

(3.68)

Fall 2013

Professor: Suchi Saria

(3.71)

Students thought that the materials covered in class were some of the better aspects of the course. Students liked learning about theory and implementation and believed the course was well balanced between the two. Students thought the workload was excessive and time consuming and that the course was unorganized. Students suggested improving the lecture quality by having the professor slow down, explain what he was writing and speak clearly about the materials. Students believed that the homework load should be spread out over the semester and considerably lightened. Prospective students should be prepared to commit a fair chunk of time to this class each week and remember to read all of the materials before the lecture. A strong understanding of math and programming is required.

Spring 2013

Professor: Suchi Saria

(4.24)

The best aspects of the course included the helpful and wel -designed homework assignments, as well as well the interesting and applicable material that students learned. The professor was very knowledgeable and showed lots of enthusiasm in teaching the subject. The worst aspects of the course included the hefty workload, and lack of guidance on the projects. The course would improve if there was a well-defined course syl abus and better discussions/assignments to help students understand what they were really learning. Prospective students should endeavor to stay on top of the work, and

Spring 2015

Professor: Suchi Saria

(3.68)

The best aspects of the course included the rich topics addressed as well as the emphasis on practical application and hands-on learning. Many students felt that delayed feedback on assignments prevented them from improving, and that lectures could have been clearer with class time used more effectively. Suggestions for improvement included shortening the homework assignments to al ow more time for the final project, and hosting review sessions more frequently. Prospective students should be able to program using Python, and should be prepared for a heavy workload.