Cloud services provider Versent have been generous in allowing the group to use some of their space for the weekly machine learning study group. The gathering has included an eclectic mix of front and back end software engineers and DevOps experts as well as being monitored by Data Scientists that do this sort of thing for breakfast.
We are now at around the half way point of Andrew Ng’s hugely popular Coursera Machine Learning which covers the essentials of machine learning and are now navigating through neural networks, having covered off linear and logistic regression with regularization to address overfitting. The format of the course includes a programming assignment which has now seen us implementing basic machine learning algorithms into Octave (kind of like a free open source version of MATLAB).
The study group has so far been a great way of drilling a bit deeper on some of the conceptual modules from experts in the field as students in our group started the course equipped with varying levels of knowledge of linear algebra, matrices and also programming. Whilst the course provides most of the programming templates, algorithm implementation can still be a bit tricky for those of us that aren’t used to writing so many loops for our day jobs. Hurrah for vectorization!
