Vowpal Wabbit is a fast, out-of-core C++ implementation of sparse gradient descent. It is sponsored by Microsoft Research and is one of the few open-source libraries capable of solving contextual bandit problems at scale—unfortunately it is also a minefield to use properly. In this talk, we’ll explore how to use Haskell to guarantee safe and performant usage of an otherwise unsafe machine learning library. Along the way, we’ll learn about some of Haskell’s lower-level language features, including the foreign function interface, inline-c, quasi-quoting & anti-quoting, synchronous & asynchronous exceptions, the bracket pattern, and some simple type-level programming.
Create a custom schedule.
Take it with you on mobile.
Get listed in the directory.
