Do you have an advanced degree in mathematics or data science from a prestigious university? Me neither! That doesn't have to put machine learning beyond our reach, though. While the underlying theory can be extremely complicated, the practical applications are very approachable for us as developers.

If you've been looking for an easy way to start dabbling in ML, this session is for you! Join me as I give an overview of ML and walk through the different types of algorithms available with some practical examples. We'll also see how we can take advantage of some of the various AWS machine learning products in order to quickly add these capabilities to our applications.


Comments are closed.

Eric Mann at 11:37 on 20 Sep 2019

Great overview of the approach to machine learning and some of the tools involved. Tooling and framework discussion could be a bit more detailed, and an illustration beyond MNIST classification would drive the points home.

Also, no PHP code at a PHP event made the code samples feel very out of place.

I really appreciated the various graph visuals as well as the walk through on how to work with bad data collection. I liked Real Estate as an easy example to follow. Machine learning has always seemed overwhelming but I feel that picked up more of an understanding in this talk. Thank you!

Kevin Boyd at 21:22 on 22 Sep 2019

Great introduction to different types of machine learning, how to apply them, and how to ensure that the data being trained against is properly prepared and analyzed. I especially liked the point about how city names shouldn't be mapped to "1, 2, 3, 4", because the ML algorithms would tend to think that the value was significant. Using columns containing 0/1 stood out for me as a solution that I wouldn't have thought of without a decent amount of research. I'm looking forward to JIT and other PHP tech that will enable this sort of tooling to be built up inside the PHP ecosystem (until then, I guess there's always cloud ML platforms ... :) ).

Bill Condo at 07:15 on 26 Sep 2019

Michael's delivery was straight forward and simple to understand. ML has been a bit of a mystery to me and this helped me start to learn and use the various strategies in future projects.