QMDP-Net: Deep Learning for Planning under Partial Observability
Abstract
This paper introduces the QMDP-net, a neural network architecture for planning under partial observability. The QMDP-net combines the strengths of model-free learning and model-based planning. It is a recurrent policy network, but it represents a policy for a parameterized set of tasks by connecting a model with a planning algorithm that solves the model, thus embedding the solution structure of planning in a network learning architecture. The QMDP-net is fully differentiable and allows for end-to-end training. We train a QMDP-net on different tasks so that it can generalize to new ones in the parameterized task set and "transfer" to other similar tasks beyond the set. In preliminary experiments, QMDP-net showed strong performance on several robotic tasks in simulation. Interestingly, while QMDP-net encodes the QMDP algorithm, it sometimes outperforms the QMDP algorithm in the experiments, as a result of end-to-end learning.
Cite
@article{arxiv.1703.06692,
title = {QMDP-Net: Deep Learning for Planning under Partial Observability},
author = {Peter Karkus and David Hsu and Wee Sun Lee},
journal= {arXiv preprint arXiv:1703.06692},
year = {2017}
}
Comments
NIPS 2017 camera-ready