Various ruminations on machine learning and mathematics
Alternative Frontends for PyMC Convex Optimized Toggle navigation Archive Publications About Alternative Frontends for PyMC Rob Zinkov 2023-11-19 When people are starting to learn pymc they often assume the syntax and workflow for the library is something that’s unchangeable. But thanks to the modular way the library is implemented, I’m going to show that it’s fairly easy to use it in a totally…
Why care about Program Synthesis Convex Optimized Toggle navigation Archive Publications About Why care about Program Synthesis Rob Zinkov 2019-02-17 Program synthesis is now emerging as an exciting new area of research not just in the programming languages community, but also the machine learning community. In this post, I’d like to convince you why this area of study has the potential to solve…
NeurIPS 2018: Papers to check out Convex Optimized Toggle navigation Archive Publications About NeurIPS 2018: Papers to check out Rob Zinkov 2018-12-21 It’s been a long time since I’ve done one of these, so below are some of the papers I found exciting at this past NeurIPS. One notable thing is as the conference has gotten larger is there are simply more papers being presented. Some people worry…
Decomposing the ELBO Convex Optimized Toggle navigation Archive Publications About Decomposing the ELBO Rob Zinkov 2018-11-02 When performing Variational Inference, we are minimizing the KL divergence between some distribution we care about \(p(\v{z} \mid \v{x})\) and some distribution that is easier to work with \(q_\phi(\v{z} \mid \v{x})\) . \[ \begin{align} \phi^* &=…
Calculating the golden-era of The Simpsons Convex Optimized Toggle navigation Archive Publications About Calculating the golden-era of The Simpsons Rob Zinkov 2017-11-03 Nathan Cunningham published last week a fantastic article about using some stats to estimate at what episode did Simpsons start to decline. Cameron Davidson-Pilon suggested this would make a great application of Bayesian…