Such convolutional neural networks are unable by their internal data representation struggle to maintain spatial hierarchies between simple and complex objects. Whereas capsule networks, which encode their data as vectors, can encode the probability of feature detection as the magnitude of the vector and the state of the detected feature in the direction of the vector. So a detected feature that moves around will have its associated vector maintain the same magnitude throughout the movement but alter their vector's orientation. Via dynamic routing, a capsule network sends lower-level capsule outputs to higher-level capsules with similar outputs--where the dot product measures similarity of vector outputs. The task of autonomous navigation is one of reinforcement learning.

In recent years, convolutional neural networks (CNN) have played an important role in the field of deep learning. Variants of CNN's have proven to be very successful in classification tasks across different domains. However, there are two big drawbacks to CNN's: their failure to take into account of important spatial hierarchies between features, and their lack of rotational invariance. As long as certain key features of an object are present in the test data, CNN's classify the test data as the object, disregarding features' relative spatial orientation to each other. This causes false positives. The lack of rotational invariance in CNN's would cause the network to incorrectly assign the object another label, causing false negatives. To address this concern, Hinton et al. propose a novel type of neural network using the concept of capsules in a recent paper. With the use of dynamic routing and reconstruction regularization, the capsule network model would be both rotation invariant and spatially aware. The capsule network has shown its potential by achieving a state-of-the-art result of 0.25% test error on MNIST without data augmentation such as rotation and scaling, better than the previous baseline of 0.39%. To further test out the application of capsule networks on data with higher dimensionality, we attempt to find the best set of configurations that yield the optimal test error on CIFAR10 dataset.