“Vowpal Wabbit” – the first time I heard this name was at a Machine Learning Meetup group in Los Angeles and the data scientist telling me about it couldn’t quite pronounce it properly and had to try several times. After the meeting I found myself muttering the name to myself for no particular reason. Very odd meme indeed! I had to find out more, so here is a top-level summary of this new machine learning platform developed originally at Yahoo! Research several years ago and currently at Microsoft Research.
A New Brand of Machine Learning
Vowpal Wabbit (aka VW) is an open source fast out-of-core learning system library and program started and led by John Langford who works at Microsoft Research New York. Vowpal Wabbit is notable as an efficient scalable implementation of online machine learning and support for a number of machine learning reductions, importance weighting, and a selection of different loss functions and optimization algorithms. The official website for the project can be found HERE and the GitHub site HERE.
Here is an introductory video for VW by John Langford presenting at the LA Machine Learning Meetup hosted by eHarmony last year. This is the group I belong to. VW was described as an ultrascale learning tool, capable of running at hardware speeds. Langford discusses two tricks of great use:
- Parallel learning over a cluster. Using this, it is easy to train on terafeature size datasets with a kilocluster.
- Learning Reductions for complex problems. Many complex prediction problems can be broken down into simple prediction problems. The reduction system in VW allows this to be done efficiently and effectively.
What’s in a Name?
“What is this undecipherable mess of vowels and consonants!?,” you ask. “That’s how Elmer Fudd would pronounce Vorpal Rabbit,” Langford answers. “Vorpal? Whatdoesthatmean?!,” you ask again. Which is where I cite the singular font of human knowledge and quote a few lines from Lewis Carrol’s Jabberwocky:
He took his vorpal sword in hand (, and later,)
One, two! One, two! And through and through
The vorpal blade went snicker-snack!
He left it dead, and with its head
He went galumphing back.
Continuing Research Success
I feel it is important to follow high-profile research projects like VW to get a pulse for the direction the field of machine learning is headed. VW deserves attention and I will continue to follow its progress and report back to you.