RLDM Conference 2017 Conference Abstract
Neural Network Memory Architectures for Autonomous Robot Navigation
- Steven Chen
- Nikolay Atanasov
- Arbaaz Khan
- Konstantinos Karydis
- Daniel Lee
- Vijay Kumar
This paper highlights the significance of including memory structures in neural networks when the latter are used to learn perception-action loops for autonomous robot navigation. Traditional navigation approaches rely on global maps of the environment to overcome cul-de-sacs and plan feasible motions. Yet, maintaining an accurate global map may be challenging in real-world settings. A possible way to mitigate this limitation is to use learning techniques that forgo hand-engineered map representations and infer appropriate control responses directly from sensed information. An important but unexplored aspect of such approaches is the effect of memory on their performance. This work is a study of memory structures for deep-neural-network-based robot navigation, and offers novel tools to train such networks from supervision and quantify their ability to generalize to unseen scenarios. We analyze the separation and generalization abilities of feedforward, long short-term memory, and differentiable neural computer networks by evaluating the generalization ability of neural networks by estimating the Vapnik-Chervonenkis (VC) dimension of maximum-margin hyperplanes trained in the feature space learned by the networks’ upstream layers. We validate that these VC-dimension measures are good predictors of actual test performance. The reported method can be applied to deep learning problems beyond robotics.