Diachronic Embedding for Temporal Knowledge Graph Completion

Abstract

Knowledge graphs (KGs) typically contain temporal facts indicating relationships among entities at different times. Due to their incompleteness, several approaches have been proposed to infer new facts for a KG based on the existing ones–a problem known as KG completion. KG embedding approaches have proved effective for KG completion, however, they have been developed mostly for static KGs. Developing temporal KG embedding models is an increasingly important problem. In this paper, we build novel models for temporal KG completion through equipping static models with a diachronic entity embedding function which provides the characteristics of entities at any point in time. This is in contrast to the existing temporal KG embedding approaches where only static entity features are provided. The proposed embedding function is model-agnostic and can be potentially combined with any static model. We prove that combining it with SimplE, a recent model for static KG embedding, results in a fully expressive model for temporal KG completion. Our experiments indicate the superiority of our proposal compared to existing baselines.

Recurrent Event Network: Autoregressive Structure Inferenceover Temporal Knowledge Graphs

“…Temporal KG Reasoning. There have been some recent attempts on incorporating temporal information in modeling dynamic knowledge graphs, broadly categorized into two settings -extrapolation (Trivedi et al, 2017) and interpolation (García-Durán et al, 2018;Leblay and Chekol, 2018;Dasgupta et al, 2018;Goel et al, 2020;Lacroix et al, 2020). For the former setting, Know-Evolve (Trivedi et al, 2017) models the occurrence of a fact as a temporal point process.…”

3DRTE: 3D Rotation Embedding in Temporal Knowledge Graph

“…García-Durán et al [12] propose a general relation-time fusion framework TA-Model with LSTM, and verify its effectiveness on some models including DistMult [4]. Rishab Goel et al [13] also focus on the design of a universal temporal fusion framework, but argue that time information should be fused with entities. With the help of Diachronic Word Embedding in NLP, they propose a method called Diachronic Entity Embedding, and prove that their model DE-Model is fully expressive.…”

“…Dataset statistics are listed in Table 1. TTransE [10] and HyTE [11] based on TransE and TransH respectively; (c) Two models with different time fusion frameworks, TA-DistMult [12] and DE-SimplE [13].…”

“…Although WIKIDATA and YAGO15K are sparse datasets with lots of long-tailed relations, their timestamp sets are smallscale. There are many facts distributed at the same time, indicates that the result is from [12], * indicates that is from [13], bold font indicates the optimal result, and underline indicates the suboptimal result; RotatE is put in a separate group because it is the only model that does not consider time. which is beneficial to the expression of time by self-attention mechanism.…”

“…Temporal Knowledge Graph (TKG) describes each time fact in the form of a quadruple: (head, relation, tail, time) and Temporal Knowledge Graph Completion (TKGC) emerges, introducing temporal characteristics to KGC. An intuitive way to realize TKGC is to integrate time information into traditional embedding models [10]- [13].…”