$K^2$-GNN: Multiple Users’ Comments Integration with Probabilistic K-Hop Knowledge Graph Neural Networks

Huixin Zhan, Kun Zhang, Chenyi Hu, Victor Sheng
Proceedings of The 13th Asian Conference on Machine Learning, PMLR 157:1477-1492, 2021.

Abstract

Integrating multiple comments into a concise statement for any online products or web services requires a non-trivial understanding of the input. Recently, graph neural networks (GNN) has been successfully applied to learn from highly-structured graph representations to mitigate the relationship between entities, such as co-references. However, current inter-sentence relation extraction cannot leverage discrete reasoning chains over multiple comments. To address this issue, in this paper, we propose a probabilistic $K$-hop knowledge graph (KKG) to extend existing knowledge graphs with inferred relations via discrete intra-sentence and inter-sentence reasoning chains. KKG associates each inferred relation with a confidence value through Bayesian inference. We further answer how a knowledge graph with inferred relations can help the multiple comments integration through integrating KKG with GNN ($\text{K}^2$-GNN). Our extensive experimental results show that our $\text{K}^2$-GNN outperforms all baseline graph models on multiple comments integration.

Cite this Paper

BibTeX
@InProceedings{pmlr-v157-zhan21b, title = {$K^2$-GNN: Multiple Users’ Comments Integration with Probabilistic K-Hop Knowledge Graph Neural Networks}, author = {Zhan, Huixin and Zhang, Kun and Hu, Chenyi and Sheng, Victor}, booktitle = {Proceedings of The 13th Asian Conference on Machine Learning}, pages = {1477--1492}, year = {2021}, editor = {Balasubramanian, Vineeth N. and Tsang, Ivor}, volume = {157}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v157/zhan21b/zhan21b.pdf}, url = {https://proceedings.mlr.press/v157/zhan21b.html}, abstract = {Integrating multiple comments into a concise statement for any online products or web services requires a non-trivial understanding of the input. Recently, graph neural networks (GNN) has been successfully applied to learn from highly-structured graph representations to mitigate the relationship between entities, such as co-references. However, current inter-sentence relation extraction cannot leverage discrete reasoning chains over multiple comments. To address this issue, in this paper, we propose a probabilistic $K$-hop knowledge graph (KKG) to extend existing knowledge graphs with inferred relations via discrete intra-sentence and inter-sentence reasoning chains. KKG associates each inferred relation with a confidence value through Bayesian inference. We further answer how a knowledge graph with inferred relations can help the multiple comments integration through integrating KKG with GNN ($\text{K}^2$-GNN). Our extensive experimental results show that our $\text{K}^2$-GNN outperforms all baseline graph models on multiple comments integration.} }
Endnote
%0 Conference Paper %T $K^2$-GNN: Multiple Users’ Comments Integration with Probabilistic K-Hop Knowledge Graph Neural Networks %A Huixin Zhan %A Kun Zhang %A Chenyi Hu %A Victor Sheng %B Proceedings of The 13th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Vineeth N. Balasubramanian %E Ivor Tsang %F pmlr-v157-zhan21b %I PMLR %P 1477--1492 %U https://proceedings.mlr.press/v157/zhan21b.html %V 157 %X Integrating multiple comments into a concise statement for any online products or web services requires a non-trivial understanding of the input. Recently, graph neural networks (GNN) has been successfully applied to learn from highly-structured graph representations to mitigate the relationship between entities, such as co-references. However, current inter-sentence relation extraction cannot leverage discrete reasoning chains over multiple comments. To address this issue, in this paper, we propose a probabilistic $K$-hop knowledge graph (KKG) to extend existing knowledge graphs with inferred relations via discrete intra-sentence and inter-sentence reasoning chains. KKG associates each inferred relation with a confidence value through Bayesian inference. We further answer how a knowledge graph with inferred relations can help the multiple comments integration through integrating KKG with GNN ($\text{K}^2$-GNN). Our extensive experimental results show that our $\text{K}^2$-GNN outperforms all baseline graph models on multiple comments integration.
APA
Zhan, H., Zhang, K., Hu, C. & Sheng, V.. (2021). $K^2$-GNN: Multiple Users’ Comments Integration with Probabilistic K-Hop Knowledge Graph Neural Networks. Proceedings of The 13th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 157:1477-1492 Available from https://proceedings.mlr.press/v157/zhan21b.html.

Related Material