MobiusE: Knowledge Graph Embedding on Mobius Ring

In this work, we propose a novel Knowledge Graph Embedding (KGE) strategy, called MobiusE, in which the entities and relations are embedded to the surface of a Mobius ring. The proposition of such a strategy is inspired by the classic TorusE, in which the addition of two arbitrary elements is subject to a modulus operation. In this sense, TorusE naturally guarantees the critical boundedness of embedding vectors in KGE. However, the nonlinear property of addition operation on Torus ring is uniquely derived by the modulus operation, which in some extent restricts the expressiveness of TorusE. As a further generalization of TorusE, MobiusE also uses modulus operation to preserve the closeness of addition on it, but the coordinates on Mobius ring interacts with each other in the following way: any vector attaches to the surface of a Mobius ring becomes its opposite one if it moves along its parametric trace by a cycle. Hence, MobiusE assumes much more nonlinear representativeness than that of TorusE, and in turn it generates much more precise embedding results. In our experiments, MobiusE outperforms TorusE and other classic embedding strategies in several key indicators. (C) 2021 Elsevier B.V. All rights reserved.

Automated summary

In this work, a novel Knowledge Graph Embedding (KGE) strategy called MobiusE is proposed, which embeds entities and relations to the surface of a Mobius ring. Compared to the classic TorusE, MobiusE exhibits more nonlinearity and generates more precise embedding results. The experiments show that MobiusE outperforms TorusE and other classic embedding strategies in several key indicators.