In the past, I wrote frequently about quadratic programming especially in R, for example here and here. It’s been a while and at least one great new library has emerged since my last post on quadratic programming — OSQP. OSQP introduces a new technique called operator splitting which offers significant performance improvements over standard interior… Read More Sparse quadratic programming with osqp
So you’ve filled an S3 bucket with hundreds of millions of objects and now, for one reason or another, you need to copy that data into another S3 bucket. How do you do that efficiently? If you’ve worked with S3 for awhile, you’ve probably seen how painfully slow it is to list all of the… Read More Copying lots of data between S3 buckets
Reading through papers on the Word2vec skip-gram model, I found myself confused on a fairly uninteresting point in the mechanics of the output layer. What was never made explicit enough (at least to me) is that the output layer returns the exact same output distribution for each context. To see why this must be true,… Read More Word2Vec: Skip-Gram Feedforward Architecture
A few months ago, I saw this post by Chris Krycho on Hackernews, which points out just how easy it can be to share Rust binaries with friends. The use case really spoke to me — I spend a lot of time building prototypes and I often need to share my prototypes with a few… Read More Rust: Cross compiling from Ubuntu to OS X
A few years ago, I wrote about how to analyze the 2012 California Health Interview Survey in R. In 2012, plans for Covered California (Obamacare in California) were just beginning to take shape. Today, Covered California is a relatively mature program and it is arguably the most successful implementation of the Affordable Care Act in… Read More Analyzing the 2015 California Health Interview Survey in R
In this post, we’ll look at a simple method to identify segments of an image based on RGB color values. The segmentation technique we’ll consider is called color quantization. Not surprisingly, this topic lends itself naturally to visualization and R makes it easy to render some really cool graphics for the color quantization problem. The… Read More Color Quantization in R
On average, Twitter users worldwide generate about 6,000 tweets per second. Obviously, there is much interest in extracting real-time signal from this rich but noisy stream of data. More generally, there many open and interesting problems in using high-velocity streaming text sources to track real-time events. In this post, I describe the key components of… Read More Building a Streaming Search Platform
This post is a follow up to my last post on quadratic programming facilities in R. A commenter pointed me to the ipoptr project which exposes an R interface to the COIN-OR optimization routine Ipopt. COIN-OR is a suite of optimization utilities implemented in C++ and supported by a back-end of configurable FORTRAN linear system… Read More Sparse Quadratic Programming with Ipoptr
This post is another tour of quadratic programming algorithms and applications in R. First, we look at the quadratic program that lies at the heart of support vector machine (SVM) classification. Then we’ll look at a very different quadratic programming demo problem that models the energy of a circus tent. The key difference between these… Read More More on Quadratic Programming in R
The California Health Interview Survey (CHIS) is a remarkable biannual survey of health status, care access, and demographics for California residents. The 2011-2012 public use survey data has recently been released and is freely available after registering at the CHIS site. CHIS currently offers data in SAS, Stata, and SPSS formats. However, thanks to a… Read More Analyzing the 2011-2012 California Health Inteview Survey with R