Workshop in English - US at php[tek] 2017
Short URL: https://joind.in/talk/fb6d3
(QR-Code (opens in new window))
Minority Report was a 2002 American science fiction film based in 2054 where police officers apprehended criminals based on foreknowledge. Machine Learning is foreknowledge; fantasy has become a reality. Attend this workshop to learn how to generate predictions you can use in your applications.
PLEASE NOTE: This half-day workshop takes place on Tuesday, May 23rd before the conference on our dedicated Workshop Day. A combination ticket including Workshop Day must be purchased in order to attend this class. Once you have purchased a Workshop Day ticket you may attend any workshops of your choice that day.
Comments
Comments are closed.
thanks! good luck with the atlanta PD data!
This isn't what I thought it would be, but it was interesting and informative just the same.
Great practical introduction to ML, and demo of AWS's ML tools. Physical handout was a nice touch!
Only complaint: workshop could touch a bit on the theory/fundamentals of ML just for a bit of context. That said, it was pretty wise to avoid describing how stochastic gradient descent actually works. :)
Solid presentation with damn good slides and handouts. This was a really really well taught course. My only concern is I would have liked to hear a little more about the ethical implications of some of the powerful tools this course was handing to students.
The example where the Chicago PD uses this tech to contact potential victims/criminals was honestly chilling. I'm worried that tool may have gotten someone hurt given how poorly some members of the Chicago PD handle community interactions. This course is definitely handing new students a loaded weapon in digital form. I want to make sure they know not to aim it at anything they don't want to see destroyed.
The idea of 'bad data in bad data out' is a good start, but digging a little more into the "no really, some of these uses could get people hurt so think before you do what your manager is asking you to do" might be good to emphasize how important it is to commit to making sure the data is clean of bias.
I know this is true for pretty much all code we do as programmers. But the AWS tools seem to make it really easy to put powerful pre-packaged kits in the hands of people without a lot of training or years of experience to develop that understanding or willingness to stand up to poorly thought out applications.