SpaceRL — A reinforcement learning-based knowledge graph driver

Abstract

Knowledge Graphs are powerful data structures used by large IT companies and the scientific community alike. They aid in the representation of related information by means of nodes connected through links indicating types of relations. These graphs are used as the basis for several smart applications, such as question answering or product recommendation. However, they are built in an automated unsupervised way, which leads to gaps in information, usually in the form of missing links between related entities in the original data source, which have to be added later by completion techniques.

SpaceRL is an end-to-end Python framework designed for the generation of reinforcement learning (RL) agents, which can be used to complete knowledge graphs through link discovery. The purpose of the generated agents is to help identify missing links in a knowledge graph by finding paths that implicitly connect two nodes, incidentally providing a reasoned explanation for the inferred new link. The generation of such agents is a complex task, even more so for a non-expert user.

SpaceRL is meant to overcome these limitations by providing a flexible set of tools designed with a wide variety of customization options, in order to adapt to different users’ needs, while also including a variety of state-of-the-art RL algorithms and several embedding models that can be combined to optimize the agents performance. Furthermore, SpaceRL offers different interfaces to make it available either locally (programmatically or via a GUI), or through an OpenAPI-compliant REST API.