This is quite a late post for me, but I just want to detail here the session I taught three weeks ago. We have a large machine learning competition in Trend Micro and I volunteered to hold a boot camp for Trend Labs, the “Philippine division”, prior to the competition.
Thinking that it was high time for Trend Labs to pursue seriously data science and specifically, machine learning, I set an agenda that ran for four hours, with a lecture and a hands-on part. At first, I didn’t know if I had complete beginners or not, so by law of averages, I prepared a slightly intimidating introduction to machine learning, complete with mathematics and jargon that full-fledged data scientists use. I wanted to have those that are slightly acquainted to have a lot of new things to consider, while on the other hand, those that are beginners can then appreciate the arduous task that lies ahead. Five minutes into the program, I got my answer. The vast majority of attendees are the uninitiated, and the minority is largely composed of my students in Deep Dive Wednesdays. That was a long day indeed.
Anyway, the first part of the competition is based on MNIST, which I call the most studied dataset in the entirety of computer vision. I put a tutorial about that in the hands-on part. I had everyone used IPython notebook and the numpy stack. Basically blitzing through an entire course on supervised learning, we managed to pull through the afternoon with surprisingly few casualties — no sleeping heads, but a lot of headaches! The passion of learning or the pressure of winning? Haha!
To everyone joining: ~30 days to go! GLHF and hope to see your team in the finals!