{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/adaptive-filters-and-aggregator-fusion-for","title":"Do We Need Anisotropic Graph Neural Networks?","arxiv_id":"2104.01481","date":"2021-04-03","proceeding":null,"authors":["Shyam A. Tailor","Felix L. Opolka","Pietro Liò","Nicholas D. Lane"],"abstract":"Common wisdom in the graph neural network (GNN) community dictates that anisotropic models -- in which messages sent between nodes are a function of both the source and target node -- are required to achieve state-of-the-art performance. Benchmarks to date have demonstrated that these models perform better than comparable isotropic models -- where messages are a function of the source node only. In this work we provide empirical evidence challenging this narrative: we propose an isotropic GNN, which we call Efficient Graph Convolution (EGC), that consistently outperforms comparable anisotropic models, including the popular GAT or PNA architectures by using spatially-varying adaptive filters. In addition to raising important questions for the GNN community, our work has significant real-world implications for efficiency. EGC achieves higher model accuracy, with lower memory consumption and latency, along with characteristics suited to accelerator implementation, while being a drop-in replacement for existing architectures. As an isotropic model, it requires memory proportional to the number of vertices in the graph ($\\mathcal{O}(V)$); in contrast, anisotropic models require memory proportional to the number of edges ($\\mathcal{O}(E)$). We demonstrate that EGC outperforms existing approaches across 6 large and diverse benchmark datasets, and conclude by discussing questions that our work raise for the community going forward. Code and pretrained models for our experiments are provided at https://github.com/shyam196/egc.","url_abs":"https://arxiv.org/abs/2104.01481v5","url_pdf":"https://arxiv.org/pdf/2104.01481v5.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":"adaptive-filters-and-aggregator-fusion-for","repo_url":"https://github.com/pyg-team/pytorch_geometric","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adaptive-filters-and-aggregator-fusion-for","repo_url":"https://github.com/shyam196/egc","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"graph-neural-network","task_name":"Graph Neural Network"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"gat","method_name":"GAT"},{"method_slug":"graph-neural-network","method_name":"Graph Neural Network"},{"method_slug":"pna","method_name":"PNA"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/graph-property-prediction-on-ogbg-code2","task":"Graph Property Prediction","dataset":"ogbg-code2","model":"EGC-M (No Edge Features)","rank_in_archive_order":12,"of":21,"metrics":{"Ext. data":"No","Number of params":"10986002","Test F1 score":"0.1595 ± 0.0019","Validation F1 score":"0.1464 ± 0.0021"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-code2","task":"Graph Property Prediction","dataset":"ogbg-code2","model":"PNA (No Edge Features)","rank_in_archive_order":15,"of":21,"metrics":{"Ext. data":"No","Number of params":"10992050","Test F1 score":"0.1570 ± 0.0032","Validation F1 score":"0.1453 ± 0.0025"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-code2","task":"Graph Property Prediction","dataset":"ogbg-code2","model":"MPNN-Max (No Edge Features)","rank_in_archive_order":17,"of":21,"metrics":{"Ext. data":"No","Number of params":"10971506","Test F1 score":"0.1552 ± 0.0022","Validation F1 score":"0.1441 ± 0.0016"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-code2","task":"Graph Property Prediction","dataset":"ogbg-code2","model":"EGC-S (No Edge Features)","rank_in_archive_order":18,"of":21,"metrics":{"Ext. data":"No","Number of params":"11156530","Test F1 score":"0.1528 ± 0.0025","Validation F1 score":"0.1427 ± 0.0020"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-molhiv","task":"Graph Property Prediction","dataset":"ogbg-molhiv","model":"EGC-M (No Edge Features)","rank_in_archive_order":32,"of":43,"metrics":{"Ext. data":"No","Number of params":"317265","Test ROC-AUC":"0.7818 ± 0.0153","Validation ROC-AUC":"0.8396 ± 0.0097"},"uses_additional_data":false},{"leaderboard":"/sota/graph-property-prediction-on-ogbg-molhiv","task":"Graph Property Prediction","dataset":"ogbg-molhiv","model":"EGC-S (No Edge Features)","rank_in_archive_order":36,"of":43,"metrics":{"Ext. data":"No","Number of params":"317013","Test ROC-AUC":"0.7721 ± 0.0110","Validation ROC-AUC":"0.8366 ± 0.0074"},"uses_additional_data":false},{"leaderboard":"/sota/node-property-prediction-on-ogbn-arxiv","task":"Node Property Prediction","dataset":"ogbn-arxiv","model":"EGC-S (100k)","rank_in_archive_order":65,"of":86,"metrics":{"Ext. data":"No","Number of params":"100648","Test Accuracy":"0.7219 ± 0.0016","Validation Accuracy":"0.7338 ± 0.0022"},"uses_additional_data":false},{"leaderboard":"/sota/node-property-prediction-on-ogbn-arxiv","task":"Node Property Prediction","dataset":"ogbn-arxiv","model":"EGC-M (100k)","rank_in_archive_order":71,"of":86,"metrics":{"Ext. data":"No","Number of params":"99464","Test Accuracy":"0.7196 ± 0.0023","Validation Accuracy":"0.7334 ± 0.0013"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2104.01481","atlas_url":"https://app.syntology.ai/?focus=2104.01481","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2104.01481"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2104.01481\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2104.01481"},"call":"get_concepts_for_paper(arxiv_id=\"2104.01481\")"}],"by_repo":[{"repo":"pyg-team/pytorch_geometric","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2606.17022"],"harvested_for_calls":[{"arxiv_id":"2606.17022","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2606.17022"},"call":"get_harvested_code_for_paper(arxiv_id=\"2606.17022\")"}]},{"repo":"shyam196/egc","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_harvested_for_is":"harvested_for: the arXiv ids (as Syntology's graph holds them) of every paper whose samples Syntology's graph links to this repository, in arXiv-id order, complete; the page names the first 3 and says how many more. A link only: it does not say the papers are the same work, that one copied the other, or that the code implements this paper's method; harvested_for_calls gives the get_harvested_code_for_paper call for each arXiv-shaped id among them","by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"1422686be7ea390c51d2f2de0eb6630513aa94e9f6c5ecf8e36cf3843f44a4c2","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}