BigSnarf blog

Infosec FTW

Monthly Archives: August 2013

Nice graphs built on d3.js

Deep Learning and InfoSec

I been reading and wishing I could start using “Deep Learning” to classify network traffic. There’s been a lot of talk over deep learning in the last year in the ML community. I recently watched Peter Norvig on the latest updates and how Google in using  deep neural networks and how they are beating conventional algorithms in their voice,  video and image classification applications.

I’m excited to hear that  malware detection, speech recognition, computer vision, and molecular activity prediction are all early adopters. I’m excited to start see network security vendors join the gaggle. The idea of neural networks is hardly new but today’s neural networks can efficiently process many more neurons, with many more layers, than before. Thanks to improvements in CPU and GPU technologies.

Geoff Hinton introduced a new algorithm  which allows for efficiently training larger and deeper neural networks than in the past.

I’m hoping that NN will provide an upper hand for InfoSec.

Simple example NetworkX IPython Notebook 1.0