Representation Learning on Graphs: A Reinforcement Learning Application
[edit]
Proceedings of Machine Learning Research, PMLR 89:33913399, 2019.
Abstract
In this work, we study value function approximation in reinforcement learning (RL) problems with high dimensional state or action spaces via a generalized version of representation policy iteration (RPI). We consider the limitations of protovalue functions (PVFs) at accurately approximating the value function in low dimensions and we highlight the importance of features learning for an improved lowdimensional value function approximation. Then, we adopt different representation learning algorithms on graphs to learn the basis functions that best represent the value function. We empirically show that node2vec, an algorithm for scalable feature learning in networks, and Graph AutoEncoder constantly outperform the commonly used smooth protovalue functions in lowdimensional feature space.
Related Material


