Best of 2018

The blog posts Hacker News loved most in 2018. 179 posts.

Math & Algorithms
125

A New Angle on L2 Regularization

Deep neural networks have been shown to be vulnerable to the adversarial example phenomenon: all models tested so far can have their classifications dramatically altered by small image perturbations . The following predictions were for instance made by a state-of-the-art network trained to recognize celebrities :

126

Just teach my kid the math

This is a new thing for me - I'm hosting a blog post by a colleague/friend - James Tanton. You can find James' original post here. One reason it's being re-hosted here is because Medium is apparently blocked in some countries: Egypt Malaysia ..... possibly others. It is astounding to me that mathematics - of all…

147

Constraint Solving with MiniZinc

This is part one of a two-part article. Part two is here. A while back somebody told me to check out MiniZinc. Eventually I did and really enjoyed it, so figured I’d write about it here. This is actually going to be a two part article. Right now we’re talking about what it is and how it works, and in the next post…