Abstract
There has been an increasing interest in inferring future links on temporal knowledge graphs (KG). While links on temporal KGs vary continuously over time, the existing approaches model the temporal KGs in discrete state spaces. To this end, we propose a novel continuum model by extending the idea of neural ordinary differential equations (ODEs) to multi-relational graph convolutional networks. The proposed model preserves the continuous nature of dynamic multi-relational graph data and encodes both temporal and structural information into continuous-time dynamic embeddings. In addition, a novel graph transition layer is applied to capture the transitions on the dynamic graph, i.e., edge formation and dissolution. We perform extensive experiments on five benchmark datasets for temporal KG reasoning, showing our model’s superior performance on the future link forecasting task.
Highlights
Most existing work (Jin et al, 2019; Zhu et al, 2020) models temporal knowledge graphs (KG) in a discrete-time domain where they take snapshots of temporal KGs sampled at regularly-spaced timestamps
LivesIn, New York) will be invalid. To this end, Inspired by neural ordinary differential equations temporal knowledge graphs were introduced. (NODEs) (Chen et al, 2018), we extend the idea
Wtrans is a trainable diagonal weight matrix and NT (o) = {(s, r)|Tsro(t) = 0)}. By employing this graph transition layer, we can better model the dynamics of temporal KGs
Summary
Inspired by (Vashishth et al, 2019) and (Yang et al, 2014), we use the entity-relation composition to model relational information. Given two graph snapshots where Emb(V, R) denotes the learnable initial em- G(t − ∆t) and G(t) at time t − ∆t and t, respecbeddings of entities and relations on the temporal tively, the graph transition tensor T(t) is defined. Wtrans is a trainable diagonal weight matrix and NT (o) = {(s, r)|Tsro(t) = 0)} By employing this graph transition layer, we can better model the dynamics of temporal KGs. We use ftrans to denote Equation 9. Score function Given the entity and relation representations at the query time tq, one can compute the scores of every triple at tq. An ablation study is conducted to show the effectiveness of our graph transition layer
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.