Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks

Physics-informed Graph Neural Networks have achieved remarkable performance in learning through graph-structured data by mitigating common GNN challenges such as over-smoothing, over-squashing, and heterophily adaption. Despite these advancements, the development of a simple yet effective paradigm t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shi, Dai, Han, Andi, Lin, Lequan, Guo, Yi, Wang, Zhiyong, Gao, Junbin
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Shi, Dai
Han, Andi
Lin, Lequan
Guo, Yi
Wang, Zhiyong
Gao, Junbin
description Physics-informed Graph Neural Networks have achieved remarkable performance in learning through graph-structured data by mitigating common GNN challenges such as over-smoothing, over-squashing, and heterophily adaption. Despite these advancements, the development of a simple yet effective paradigm that appropriately integrates previous methods for handling all these challenges is still underway. In this paper, we draw an analogy between the propagation of GNNs and particle systems in physics, proposing a model-agnostic enhancement framework. This framework enriches the graph structure by introducing additional nodes and rewiring connections with both positive and negative weights, guided by node labeling information. We theoretically verify that GNNs enhanced through our approach can effectively circumvent the over-smoothing issue and exhibit robustness against over-squashing. Moreover, we conduct a spectral analysis on the rewired graph to demonstrate that the corresponding GNNs can fit both homophilic and heterophilic graphs. Empirical validations on benchmarks for homophilic, heterophilic graphs, and long-term graph datasets show that GNNs enhanced by our method significantly outperform their original counterparts.
doi_str_mv 10.48550/arxiv.2401.14580
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2401_14580</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2401_14580</sourcerecordid><originalsourceid>FETCH-LOGICAL-a670-6f20dca9ae1791d1e08fc5c5720bd0e36cff0e6ed95684ee9568d11b268d3f663</originalsourceid><addsrcrecordid>eNotz0FPwjAYxvFePBj0A3ji_QKb7bZ2m7cFEUlQOODB0yzt260ROtIOkG_vQE-_5Dk8yZ-QB0bjrOCcPkr_Y49xklEWs4wX9JZ8PWOwjYPP7uBheXLw4ewRfcAnqGDVnoNVIZo70_kdaqga14XeKnjDvu00DDNMXSudsq6BmZf7Ft7x4OV2oD91_jvckRsjtwHv_x2R9ct0PXmNFsvZfFItIilyGgmTUK1kKZHlJdMMaWEUVzxP6EZTTIUyhqJAXXJRZIgXNGObZCA1QqQjMv67vSbWe2930p_rS2p9TU1_AYJjTzc</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks</title><source>arXiv.org</source><creator>Shi, Dai ; Han, Andi ; Lin, Lequan ; Guo, Yi ; Wang, Zhiyong ; Gao, Junbin</creator><creatorcontrib>Shi, Dai ; Han, Andi ; Lin, Lequan ; Guo, Yi ; Wang, Zhiyong ; Gao, Junbin</creatorcontrib><description>Physics-informed Graph Neural Networks have achieved remarkable performance in learning through graph-structured data by mitigating common GNN challenges such as over-smoothing, over-squashing, and heterophily adaption. Despite these advancements, the development of a simple yet effective paradigm that appropriately integrates previous methods for handling all these challenges is still underway. In this paper, we draw an analogy between the propagation of GNNs and particle systems in physics, proposing a model-agnostic enhancement framework. This framework enriches the graph structure by introducing additional nodes and rewiring connections with both positive and negative weights, guided by node labeling information. We theoretically verify that GNNs enhanced through our approach can effectively circumvent the over-smoothing issue and exhibit robustness against over-squashing. Moreover, we conduct a spectral analysis on the rewired graph to demonstrate that the corresponding GNNs can fit both homophilic and heterophilic graphs. Empirical validations on benchmarks for homophilic, heterophilic graphs, and long-term graph datasets show that GNNs enhanced by our method significantly outperform their original counterparts.</description><identifier>DOI: 10.48550/arxiv.2401.14580</identifier><language>eng</language><subject>Computer Science - Learning</subject><creationdate>2024-01</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,781,886</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2401.14580$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2401.14580$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Shi, Dai</creatorcontrib><creatorcontrib>Han, Andi</creatorcontrib><creatorcontrib>Lin, Lequan</creatorcontrib><creatorcontrib>Guo, Yi</creatorcontrib><creatorcontrib>Wang, Zhiyong</creatorcontrib><creatorcontrib>Gao, Junbin</creatorcontrib><title>Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks</title><description>Physics-informed Graph Neural Networks have achieved remarkable performance in learning through graph-structured data by mitigating common GNN challenges such as over-smoothing, over-squashing, and heterophily adaption. Despite these advancements, the development of a simple yet effective paradigm that appropriately integrates previous methods for handling all these challenges is still underway. In this paper, we draw an analogy between the propagation of GNNs and particle systems in physics, proposing a model-agnostic enhancement framework. This framework enriches the graph structure by introducing additional nodes and rewiring connections with both positive and negative weights, guided by node labeling information. We theoretically verify that GNNs enhanced through our approach can effectively circumvent the over-smoothing issue and exhibit robustness against over-squashing. Moreover, we conduct a spectral analysis on the rewired graph to demonstrate that the corresponding GNNs can fit both homophilic and heterophilic graphs. Empirical validations on benchmarks for homophilic, heterophilic graphs, and long-term graph datasets show that GNNs enhanced by our method significantly outperform their original counterparts.</description><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz0FPwjAYxvFePBj0A3ji_QKb7bZ2m7cFEUlQOODB0yzt260ROtIOkG_vQE-_5Dk8yZ-QB0bjrOCcPkr_Y49xklEWs4wX9JZ8PWOwjYPP7uBheXLw4ewRfcAnqGDVnoNVIZo70_kdaqga14XeKnjDvu00DDNMXSudsq6BmZf7Ft7x4OV2oD91_jvckRsjtwHv_x2R9ct0PXmNFsvZfFItIilyGgmTUK1kKZHlJdMMaWEUVzxP6EZTTIUyhqJAXXJRZIgXNGObZCA1QqQjMv67vSbWe2930p_rS2p9TU1_AYJjTzc</recordid><startdate>20240125</startdate><enddate>20240125</enddate><creator>Shi, Dai</creator><creator>Han, Andi</creator><creator>Lin, Lequan</creator><creator>Guo, Yi</creator><creator>Wang, Zhiyong</creator><creator>Gao, Junbin</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240125</creationdate><title>Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks</title><author>Shi, Dai ; Han, Andi ; Lin, Lequan ; Guo, Yi ; Wang, Zhiyong ; Gao, Junbin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a670-6f20dca9ae1791d1e08fc5c5720bd0e36cff0e6ed95684ee9568d11b268d3f663</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Shi, Dai</creatorcontrib><creatorcontrib>Han, Andi</creatorcontrib><creatorcontrib>Lin, Lequan</creatorcontrib><creatorcontrib>Guo, Yi</creatorcontrib><creatorcontrib>Wang, Zhiyong</creatorcontrib><creatorcontrib>Gao, Junbin</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Shi, Dai</au><au>Han, Andi</au><au>Lin, Lequan</au><au>Guo, Yi</au><au>Wang, Zhiyong</au><au>Gao, Junbin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks</atitle><date>2024-01-25</date><risdate>2024</risdate><abstract>Physics-informed Graph Neural Networks have achieved remarkable performance in learning through graph-structured data by mitigating common GNN challenges such as over-smoothing, over-squashing, and heterophily adaption. Despite these advancements, the development of a simple yet effective paradigm that appropriately integrates previous methods for handling all these challenges is still underway. In this paper, we draw an analogy between the propagation of GNNs and particle systems in physics, proposing a model-agnostic enhancement framework. This framework enriches the graph structure by introducing additional nodes and rewiring connections with both positive and negative weights, guided by node labeling information. We theoretically verify that GNNs enhanced through our approach can effectively circumvent the over-smoothing issue and exhibit robustness against over-squashing. Moreover, we conduct a spectral analysis on the rewired graph to demonstrate that the corresponding GNNs can fit both homophilic and heterophilic graphs. Empirical validations on benchmarks for homophilic, heterophilic graphs, and long-term graph datasets show that GNNs enhanced by our method significantly outperform their original counterparts.</abstract><doi>10.48550/arxiv.2401.14580</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2401.14580
ispartof
issn
language eng
recordid cdi_arxiv_primary_2401_14580
source arXiv.org
subjects Computer Science - Learning
title Design Your Own Universe: A Physics-Informed Agnostic Method for Enhancing Graph Neural Networks
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-15T15%3A22%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Design%20Your%20Own%20Universe:%20A%20Physics-Informed%20Agnostic%20Method%20for%20Enhancing%20Graph%20Neural%20Networks&rft.au=Shi,%20Dai&rft.date=2024-01-25&rft_id=info:doi/10.48550/arxiv.2401.14580&rft_dat=%3Carxiv_GOX%3E2401_14580%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true