DHMConv: Directed Hypergraph Momentum Convolution Framework

Wenbo Zhao, Zitong Ma, Zhe Yang
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, PMLR 238:3385-3393, 2024.

Abstract

Due to its capability to capture high-order information, the hypergraph model has shown greater potential than the graph model in various scenarios. Real-world entity relations frequently involve directionality, in order to express high-order information while capturing directional information in relationships, we present a directed hypergraph spatial convolution framework that is designed to acquire vertex embeddings of directed hypergraphs. The framework characterizes the information propagation of directed hypergraphs through two stages: hyperedge information aggregation and hyperedge information broadcasting. During the hyperedge information aggregation stage, we optimize the acquisition of hyperedge information using attention mechanisms. In the hyperedge information broadcasting stage, we leverage a directed hypergraph momentum encoder to capture the directional information of directed hyperedges. Experimental results on five publicly available directed graph datasets of three different categories demonstrate that our proposed DHMConv outperforms various commonly used graph and hypergraph models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v238-zhao24c, title = { {DHMConv}: Directed Hypergraph Momentum Convolution Framework }, author = {Zhao, Wenbo and Ma, Zitong and Yang, Zhe}, booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, pages = {3385--3393}, year = {2024}, editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, volume = {238}, series = {Proceedings of Machine Learning Research}, month = {02--04 May}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v238/zhao24c/zhao24c.pdf}, url = {https://proceedings.mlr.press/v238/zhao24c.html}, abstract = { Due to its capability to capture high-order information, the hypergraph model has shown greater potential than the graph model in various scenarios. Real-world entity relations frequently involve directionality, in order to express high-order information while capturing directional information in relationships, we present a directed hypergraph spatial convolution framework that is designed to acquire vertex embeddings of directed hypergraphs. The framework characterizes the information propagation of directed hypergraphs through two stages: hyperedge information aggregation and hyperedge information broadcasting. During the hyperedge information aggregation stage, we optimize the acquisition of hyperedge information using attention mechanisms. In the hyperedge information broadcasting stage, we leverage a directed hypergraph momentum encoder to capture the directional information of directed hyperedges. Experimental results on five publicly available directed graph datasets of three different categories demonstrate that our proposed DHMConv outperforms various commonly used graph and hypergraph models. } }
Endnote
%0 Conference Paper %T DHMConv: Directed Hypergraph Momentum Convolution Framework %A Wenbo Zhao %A Zitong Ma %A Zhe Yang %B Proceedings of The 27th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2024 %E Sanjoy Dasgupta %E Stephan Mandt %E Yingzhen Li %F pmlr-v238-zhao24c %I PMLR %P 3385--3393 %U https://proceedings.mlr.press/v238/zhao24c.html %V 238 %X Due to its capability to capture high-order information, the hypergraph model has shown greater potential than the graph model in various scenarios. Real-world entity relations frequently involve directionality, in order to express high-order information while capturing directional information in relationships, we present a directed hypergraph spatial convolution framework that is designed to acquire vertex embeddings of directed hypergraphs. The framework characterizes the information propagation of directed hypergraphs through two stages: hyperedge information aggregation and hyperedge information broadcasting. During the hyperedge information aggregation stage, we optimize the acquisition of hyperedge information using attention mechanisms. In the hyperedge information broadcasting stage, we leverage a directed hypergraph momentum encoder to capture the directional information of directed hyperedges. Experimental results on five publicly available directed graph datasets of three different categories demonstrate that our proposed DHMConv outperforms various commonly used graph and hypergraph models.
APA
Zhao, W., Ma, Z. & Yang, Z.. (2024). DHMConv: Directed Hypergraph Momentum Convolution Framework . Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:3385-3393 Available from https://proceedings.mlr.press/v238/zhao24c.html.

Related Material