Alternatives to a Degree to Prove Yourself in Deep Learning

This post has been translated into Chinese here.

This week’s Ask-A-Data-Scientist column has a question from a college freshman at my alma mater, Swarthmore. Please email your data science related quandaries to [email protected]. Note that questions are edited for clarity and brevity. Previous posts include:

Q: I’m currently a freshman at Swarthmore College and I’m really interested in machine learning and deep learning. I wanted to take Artificial Intelligence this semester; unfortunately, no freshmen got into the class as it has been difficult for the CS department to keep up with the huge spike in interest.

I’m currently taking Andrew Ng’s Coursera Course on Machine Learning and will finish it in ~2-3 weeks. Next, I was planning on taking your MOOC, which I saw on hacker news.

I know you may be too busy, but can I ask you questions I have about ML and my proposed plan? How can I continue to learn machine learning after Ng’s Coursera course and It seems like the only two options are 1.) research and 2.) graduate level courses at UPenn (which seem to be quite difficult to get into from Swarthmore (especially as a first-year student)). Any advice would be appreciated.

A: In general, I am happy to answer questions, although it may take me some time (my inbox, oh my inbox). For technical questions, it’s best to first ask on our forums. There are tons of interesting discussions on our forums, even if you are not taking our course. For career-related or general questions, I often answer them in my ask-a-data-scientist column.

  1. Even without Swarthmore or UPenn’s AI classes, you will never run out of things to do with deep learning or ways to learn more. Our MOOC takes 70 hours of study to complete, and if you get interested in any of the Kaggle competitions we have you start, you could spend much longer. We will be releasing Part 2 in a few months, which will be a similar time commitment, only with even more side avenues for further study, recommended papers to read, and ways to extend the work.

  2. Take the official classes when/if you are able, but you don’t need the credentials or resources from official classes (to anyone out there not in university or at a university that doesn’t offer an AI class, don’t worry: you don’t need them!). One of our students, who was an econ major with no graduate degree, was just accepted to the prestigious Google Brain residency program! Another student developed a new fraud detection technique based on material from our course and has received a bonus at his job. Several others have received internship and job offers, or switched teams in their current workplaces to more exciting machine learning projects.

Credentials can sometimes be useful to get your foot in the door, particularly if you are an underrepresented minority in tech (and thus facing greater scrutiny).

However, there are lots of even more effective ways to get your name and work out there:

In general, I recommend that you start a side project of something that interests you (that uses deep learning) so you will have that to work on.

Why you (yes, you) should blog

The top advice I would give my younger self would be to start blogging sooner. Here are some reasons to blog:

To inspire you, here are some sample blog posts from students in part 2 of our course:

I enjoyed all of the above blog posts and also, I don’t think any of them are too intimidating. They’re meant to be accessible.

Tips for getting started blogging

Jeremy had been suggesting for years that I should start blogging, and I’d respond “I don’t have anything to say.” This wasn’t true. What I meant was that I didn’t feel confident, and I felt like the things I could write had already been written about by people with more expertise or better writing skills than me.

It turns out that is fine! Your posts don’t have to be earth-shattering or even novel to be read and shared. My writing skills were rather weak when I started (part of the reason I chose to study math and CS in college was because those courses requried the least amount of writing and also no labs), but my skills are improving with time.

Here are some more tips to help you start your first post:

You are on the right path by taking MOOCs, and by adding in a side project, involvement in online communities, and blogging you will have even more opportunities to learn and meet others!