My latest work on deep learning security draws inspiration from the temporal computation found in spiking neural networks (SNN) to make having secure networks more accessible. State of the art techniques to train deep networks that are robust to adversarial examples - small, crafted perturbations to input images that fool networks with high confidence - are very computationally expensive and often require redundant neural networks and the injection of many additional training examples. Temporal computation promotes decision-making in deep neural networks that is more context driven and makes feature extraction more robust to small adversarial perturbations.