My presentations

  • Dynamic Routing Between Capsules


    Paper presentation of 'Dynamic Routing Between Capsules'. Capsules are groups of neurons that contain extra information about an entity's size, pose, location, etc. than typically capable in a deep convnet. This paper demonstrates how a shallow capsule network can be used in conjuction with 'routing by agreement' to acheive near-state of the art in image classification tasks.
  • Differentiable Neural Computers


    Created in 2016 as an extension of the Neural Turing Machine by Alex Graves and Greg Wayne (and others), Differentiable Neural Computers a type of memory-augmented deep neural networks that can be trained by gradient descent.
  • Building and Breaking Deep Neural Networks with TensorFlow


    Learn the basics of TensorFlow, how to build an MNIST classifier, and then fool that classifier with adversarial examples.
  • vim and tmux: Edit at the Speed of Thought


    Rule the command line with this introduction to editing with vim and manipulating terminal sessions with tmux.