{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/edge-augmented-graph-transformers-global-self","title":"Global Self-Attention as a Replacement for Graph Convolution","arxiv_id":"2108.03348","date":"2021-08-07","proceeding":null,"authors":["Md Shamim Hussain","Mohammed J. Zaki","Dharmashankar Subramanian"],"abstract":"We propose an extension to the transformer neural network architecture for general-purpose graph learning by adding a dedicated pathway for pairwise structural information, called edge channels. The resultant framework - which we call Edge-augmented Graph Transformer (EGT) - can directly accept, process and output structural information of arbitrary form, which is important for effective learning on graph-structured data. Our model exclusively uses global self-attention as an aggregation mechanism rather than static localized convolutional aggregation. This allows for unconstrained long-range dynamic interactions between nodes. Moreover, the edge channels allow the structural information to evolve from layer to layer, and prediction tasks on edges/links can be performed directly from the output embeddings of these channels. We verify the performance of EGT in a wide range of graph-learning experiments on benchmark datasets, in which it outperforms Convolutional/Message-Passing Graph Neural Networks. EGT sets a new state-of-the-art for the quantum-chemical regression task on the OGB-LSC PCQM4Mv2 dataset containing 3.8 million molecular graphs. Our findings indicate that global self-attention based aggregation can serve as a flexible, adaptive and effective replacement of graph convolution for general-purpose graph learning. Therefore, convolutional local neighborhood aggregation is not an essential inductive bias.","url_abs":"https://arxiv.org/abs/2108.03348v3","url_pdf":"https://arxiv.org/pdf/2108.03348v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"edge-augmented-graph-transformers-global-self","repo_url":"https://github.com/shamim-hussain/egt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"edge-augmented-graph-transformers-global-self","repo_url":"https://github.com/shamim-hussain/egt_pytorch","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"edge-augmented-graph-transformers-global-self","repo_url":"https://github.com/shamim-hussain/egt_triangular","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"edge-classification","task_name":"Edge Classification"},{"task_slug":"graph-classification","task_name":"Graph Classification"},{"task_slug":"graph-learning","task_name":"Graph Learning"},{"task_slug":"graph-property-prediction","task_name":"Graph Property Prediction"},{"task_slug":"graph-regression","task_name":"Graph Regression"},{"task_slug":"inductive-bias","task_name":"Inductive Bias"},{"task_slug":"link-prediction","task_name":"Link Prediction"},{"task_slug":"node-classification","task_name":"Node Classification"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"egt","method_name":"EGT"},{"method_slug":"graph-convolutional-networks","method_name":"Graph Convolutional Networks"},{"method_slug":"graph-transformer","method_name":"Graph Transformer"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"lapeigen","method_name":"LapEigen"},{"method_slug":"laplacian-pe","method_name":"Laplacian PE"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[{"slug":"egt","name":"EGT","full_name":"Edge-augmented Graph Transformer"}],"results":[{"leaderboard":"/sota/graph-classification-on-cifar10-100k","task":"Graph Classification","dataset":"CIFAR10 100k","model":"EGT","rank_in_archive_order":15,"of":20,"metrics":{"Accuracy (%)":"68.702"},"uses_additional_data":false},{"leaderboard":"/sota/graph-classification-on-mnist","task":"Graph Classification","dataset":"MNIST","model":"EGT","rank_in_archive_order":10,"of":13,"metrics":{"Accuracy":"98.173"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-molhiv","task":"Graph Property Prediction","dataset":"ogbg-molhiv","model":"EGT","rank_in_archive_order":13,"of":43,"metrics":{"Test ROC-AUC":"0.806 ± 0.0065"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-molpcba","task":"Graph Property Prediction","dataset":"ogbg-molpcba","model":"EGT","rank_in_archive_order":16,"of":36,"metrics":{"Test AP":"0.2961 ± 0.0024"},"uses_additional_data":false},{"leaderboard":"/sota/graph-regression-on-pcqm4m-lsc","task":"Graph Regression","dataset":"PCQM4M-LSC","model":"EGT","rank_in_archive_order":10,"of":11,"metrics":{"Validation MAE":"0.1224"},"uses_additional_data":false},{"leaderboard":"/sota/graph-regression-on-pcqm4mv2-lsc","task":"Graph Regression","dataset":"PCQM4Mv2-LSC","model":"EGT + Triangular Attention","rank_in_archive_order":2,"of":20,"metrics":{"Test MAE":"0.0683","Validation MAE":"0.0671"},"uses_additional_data":false},{"leaderboard":"/sota/graph-regression-on-pcqm4mv2-lsc","task":"Graph Regression","dataset":"PCQM4Mv2-LSC","model":"EGT","rank_in_archive_order":10,"of":20,"metrics":{"Test MAE":"0.0862","Validation MAE":"0.0857"},"uses_additional_data":false},{"leaderboard":"/sota/graph-regression-on-zinc-100k","task":"Graph Regression","dataset":"ZINC 100k","model":"EGT","rank_in_archive_order":4,"of":8,"metrics":{"MAE":"0.143"},"uses_additional_data":false},{"leaderboard":"/sota/graph-regression-on-zinc-500k","task":"Graph Regression","dataset":"ZINC-500k","model":"EGT","rank_in_archive_order":24,"of":36,"metrics":{"MAE":"0.108"},"uses_additional_data":false},{"leaderboard":"/sota/link-prediction-on-tsp-hcp-benchmark-set","task":"Link Prediction","dataset":"TSP/HCP Benchmark set","model":"EGT","rank_in_archive_order":3,"of":4,"metrics":{"F1":"0.853"},"uses_additional_data":false},{"leaderboard":"/sota/node-classification-on-cluster","task":"Node Classification","dataset":"CLUSTER","model":"EGT","rank_in_archive_order":2,"of":12,"metrics":{"Accuracy":"79.232"},"uses_additional_data":false},{"leaderboard":"/sota/node-classification-on-pattern","task":"Node Classification","dataset":"PATTERN","model":"EGT","rank_in_archive_order":5,"of":11,"metrics":{"Accuracy":"86.821"},"uses_additional_data":false},{"leaderboard":"/sota/node-classification-on-pattern-100k","task":"Node Classification","dataset":"PATTERN 100k","model":"EGT","rank_in_archive_order":1,"of":9,"metrics":{"Accuracy (%)":"86.816"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2108.03348","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2108.03348"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/shamim-hussain/egt_triangular","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/shamim-hussain/egt_pytorch","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/shamim-hussain/egt","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":4},"by_repo_kind":{"official":{"samples":4,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"23012f87aaa39c42","entry":"import_object","repo":"shamim-hussain/egt","repo_kind":"official","path":"lib/training/importer.py","file_url":"https://github.com/shamim-hussain/egt/blob/HEAD/lib/training/importer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"23012f87aaa39c42"}},{"code_sha256_prefix":"ca5300d8a50237fe","entry":"import_scheme","repo":"shamim-hussain/egt","repo_kind":"official","path":"lib/training/importer.py","file_url":"https://github.com/shamim-hussain/egt/blob/HEAD/lib/training/importer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ca5300d8a50237fe"}},{"code_sha256_prefix":"4b0f182cd45413d3","entry":"read_config_from_file","repo":"shamim-hussain/egt","repo_kind":"official","path":"lib/training/training_base.py","file_url":"https://github.com/shamim-hussain/egt/blob/HEAD/lib/training/training_base.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4b0f182cd45413d3"}},{"code_sha256_prefix":"219b23ce1676877d","entry":"save_config_to_file","repo":"shamim-hussain/egt","repo_kind":"official","path":"lib/training/training_base.py","file_url":"https://github.com/shamim-hussain/egt/blob/HEAD/lib/training/training_base.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"219b23ce1676877d"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}