End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion

By Mingmin Zhen, Jinglu Wang, Lei Zhou, Tian Fang, Long Quan. AAAI Conference on Artificial Intelligence, 2018

732 citations recorded by the source corpus.

Knowledge graph embedding has been an active research topic for knowledge base completion, with progressive improvement from the initial TransE, TransH, DistMult et al to the current state-of-the-art ConvE. ConvE uses 2D convolution over embeddings and multiple layers of nonlinear features to model knowledge graphs. The model can be efficiently trained and scalable to large knowledge graphs. However, there is no structure enforcement in the embedding space of ConvE. The recent graph convolutional network (GCN) provides another way of learning graph node embedding by successfully utilizing graph connectivity structure. In this work, we propose a novel end-to-end Structure-Aware Convolutional Network (SACN) that takes the benefit of GCN and ConvE together. SACN consists of an encoder of a weighted graph convolutional network (WGCN), and a decoder of a convolutional network called Conv-TransE. WGCN utilizes knowledge graph node structure, node attributes and edge relation types. It has learnable weights that adapt the amount of information from neighbors used in local aggregation, leading to more accurate embeddings of graph nodes. Node attributes in the graph are represented as additional nodes in the WGCN. The decoder Conv-TransE enables the state-of-the-art ConvE to be translational between entities and relations while keeps the same link prediction performance as ConvE. We demonstrate the effectiveness of the proposed SACN on standard FB15k-237 and WN18RR datasets, and it gives about 10% relative improvement over the state-of-the-art ConvE in terms of HITS@1, HITS@3 and HITS@10.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/8f3812a4f2093b488fa4bd3adda3cea0e770dc3d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Temporal Knowledge Graph Reasoning Based on Evolutional Representation Learning (2021, methodology): “In more detail, we select ConvE [6], ConvTransE [30] from the static models, as well as RGCRN [29] from the temporal models.”
A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal (2022, background): “2022 RED-GNN [82] Inductive GNN 2019 SACN [61] Transductive GNN”
A Survey on Knowledge Graphs: Representation, Acquisition, and Applications (2020, background): “Its scoring function is defined as fr (h, t) =W ×1 h×2 r×3 t, (26) 7 h r t W W W tanh w (a) MLP Convolution Features maps * hrt Score Filters (b) CNN ··· GCN Activation Entity/Relation Embedding KG (c) GCN combine combine RNN unit e1 e2 e3r1 r2 (d) RSN Fig.”
Neural, symbolic and neural-symbolic reasoning on knowledge graphs (2021, methodology): “Instead of using DisMult as the decoder, SACN [122] uses a variant”
Knowledge Graph Completion: A Review (2020, background): “Shang et al. combined the characteristics of GCN and ConvE model and proposed an end-to-end Structure-Aware Convolutional Network (SACN) [94] to complete the knowledge graph.”
InteractE: Improving Convolution-based Knowledge Graph Embeddings by Increasing Feature Interactions (2019, methodology): “The scores of all the baselines are taken directly from the values reported in the papers (Dettmers et al. 2018; Sun et al. 2019a; Shang et al. 2019; Cai and Wang 2018; García-Durán and Niepert 2018).”
Composition-based Multi-Relational Graph Convolutional Networks (2019, methodology): “Hence, we extend E and R with corresponding inverse edges and relations, i.e., O ( Kd 2 ) Weighted-GCN Shang et al. (2019) O ( Kd 2 + K |R| ) Relational-GCN Schlichtkrull et al. (2017) O ( B Kd 2 + B K |R| ) Representing relations as vectors alleviates the problem of over-parameterization while…”
Large-scale Multi-modal Pre-trained Models: A Comprehensive Survey (2023, unknown): “A structure-aware convolutional network (SACN) takes the benefit of GCN and ConvE [213] together, where GCN as the encoder utilizes knowledge graph node structure and ConvE as the decoder enables the translational feature [214].”
A review on the long short-term memory model (2020, background): “In the context of predicting future frames, Fan et al. (2019) introduced cubic LSTM consisting of three branches, namely a spatial, a temporal and an output branch, the latter of which combines the ﬁrst two branches to generate the predicted frames.”