Hacking the MOOC

I’m in the middle of Week 6 of the Andrew Ng Machine Learning Coursera Course. Prior to starting the course I wrote the following about sucessfully completing the course. Spoiler alert: the road has been rocky.

I have attempted the Machine Learning Coursera Course at least 3 times. I have yet to once complete the course or more than 3 programming assignments. In fact, I have signed up for and attempted a number of other Coursera courses without success. My hope was that trying shorter courses would help me complete a course, but in all honesty, the motivation has been lacking. No doubt I have learned a ton even with these failures, but I really would like to complete a course.

What are my MOTIVATIONS?

Despite my failures I still have completing the course as goal. I want to complete a Coursera course for a number of reasons:

  1. “Hack the MOOC”: The wealth of information and knowledge and training available in Coursera, Udacity and other MOOCs is huge. As a software developer, who is constantly striving to teach herself new things, I want to figure out how to fit MOOCs into my busy life.
  2. MOOC as an occasion for Deliberate Practice: We’ve all read the pop psych books hailing the boons of deliberate practice. By working on tasks that challenge my specific deficiencies, we can grow our skill sets to (hopefully) reach mastery.
  3. Be able to keep up with a major shift in how we as professionals learn. And hopefully, in figuring this out, share with others how to do the same.
  4. Love of Learning: I was the girl who completed the unfinished chapters from school textbooks over the summer. I taught myself pre-algebra after my mother asked me if they were teaching me about “x” in school. What was this “x”? I still have all of my GRE vocabulary hand-written flash cards and study materials. I love the process of learning more that I love the knowledge itself – I think sometimes.

Specifically, I want to complete the Andrew Ng for the following reason’s:

  1. My “Workie Work”: I work for a retailer and have worked with Hadoop and Cassandra since 2011. Advanced analytics is an area of interest to me for professional reasons.
  2. This is a topic that has fascinated me since my undergraduate degree. My senior thesis in college was on Neural Networks. The though of training a computer to derive answers to questions that I myself could not discern is fascinating. Think of all of the NEW information we could discern about the world to help others using computers!
  3. Some specific things I’m interested in building using my knowledge from the class: predict which of my pins will be repinned, predict how much water my growerbot will need.
  4. Autonomous Flying Helicopters…. omgbbqwtf, need I say more.

Confidence that this work is at my skill level.

How I plan to receive FEEDBACK?

  • Start a meetup to discuss the Course with the group:

What is my plan for encouraging REPETITION?

  • Pomodoro

  • Blogging

  • Taking notes by hand on a Nice New Notebook

  • The Coursera app: strategies for good usage (other apps)

  • Understanding my weaknesses and inventing specific tasks to address these deficiencies.

Prepping before hand: Getting my 9 to 5 Work out of the way.

The interesting part, now that I am half way through the course, is that I anticipated difficulties in maintaining concentration and designed several strategies for avoiding loosing concentration and for increasing my productivity while studying.

The actuality of my experience has been that the single most difficult part of completing the course so far has been in making time to get through all of the videos, quizzes and homework. Establishing boundaries in my work and personal schedule has been difficult. Late night production support, my back going out twice and being unable to exercise and barely get around, and weight gain have been just a few of the issues that I’ve run into as a result of trying to complete this course by not establishing time boundaries. Without these boundaries in place, implementing any of these strategies is near impossible.

So, my goal for the second half of the course is to look at “How to get my 9 to 5 work out of the way and set aside a consistent time daily for working on the Machine Learning Coursera Class?”.

The Bright Spots have been in taking large chunks of time for the homework assignments, working remotely on my work machine.

Attempt at least some of the homework prior to the Meetup on Tuesday evenings, which means that videos and quizzes should be completed prior to Tuesday.

Taking one hour from my work day to focus on watching the course videos.

Don’t plan work for after 6PM. If there is a need for work after 6PM, request resources to work on it in the evening. Don’t commit to any timelines that require writing code after 5PM.

So, here’s some things that I would like to try in the second half of the course:

  • Try some of the techniques that I was not able to attempt during the first half of the course like blogging and doing pomodoro’s for videos and homework. Who knows? Those might be very useful strategies.

  • Set a time everyday for working on the class.

  • I’ve thought about using Beeminder for tracking how much time I spend on the course.

  • Precommit to studying everyday by telling everyone that I know that my goal is to study everyday. Other possibilities for precommitment are: tweet, facebook post, blog.

  • Another thing that I’ve been wanting to try is just to experiment with some different learning techniques to see what helps make the learning process more enjoyable.

  • Set a goal of having a working prototype of a Machine Learning problem that I am interested in.

I have also been recently admitted to gradute school in computer science, and I think of this MOOC as a dry run of my first semester.