{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,1]],"date-time":"2026-08-01T18:32:27Z","timestamp":1785609147037,"version":"3.56.0"},"reference-count":22,"publisher":"Springer Science and Business Media LLC","issue":"7","license":[{"start":{"date-parts":[[2025,4,7]],"date-time":"2025-04-07T00:00:00Z","timestamp":1743984000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,4,7]],"date-time":"2025-04-07T00:00:00Z","timestamp":1743984000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100013209","name":"Hellenic Foundation for Research and Innovation","doi-asserted-by":"publisher","award":["10860"],"award-info":[{"award-number":["10860"]}],"id":[{"id":"10.13039\/501100013209","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Appl Intell"],"published-print":{"date-parts":[[2025,5]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p>In this work, we generalize the ideas of Kaiming initialization to Graph Neural Networks (GNNs) and propose a new scheme (G-Init) that reduces oversmoothing, leading to very good results in node and graph classification tasks. GNNs are commonly initialized using methods designed for other types of Neural Networks, overlooking the underlying graph topology. We analyze theoretically the variance of signals flowing forward and gradients flowing backward in the class of convolutional GNNs. We then simplify our analysis to the case of the GCN and propose a new initialization method. Results indicate that the new method (G-Init) reduces oversmoothing in deep GNNs, facilitating their effective use. Our approach achieves an accuracy of 61.60% on the <jats:italic>CS<\/jats:italic> dataset (32-layer GCN) and 69.24% on <jats:italic>Cora<\/jats:italic> (64-layer GCN), surpassing state-of-the-art initialization methods by 25.6 and 8.6 percentage points, respectively. Extensive experiments confirm the robustness of our method across multiple benchmark datasets, highlighting its effectiveness in diverse settings. Furthermore, our experimental results support the theoretical findings, demonstrating the advantages of deep networks in scenarios with no feature information for unlabeled nodes (i.e., \u201ccold start\u201d scenario).<\/jats:p>","DOI":"10.1007\/s10489-025-06426-0","type":"journal-article","created":{"date-parts":[[2025,4,7]],"date-time":"2025-04-07T04:25:23Z","timestamp":1743999923000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":8,"title":["Reducing oversmoothing through informed weight initialization in graph neural networks"],"prefix":"10.1007","volume":"55","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3434-2717","authenticated-orcid":false,"given":"Dimitrios","family":"Kelesis","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dimitris","family":"Fotakis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Georgios","family":"Paliouras","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,4,7]]},"reference":[{"key":"6426_CR1","unstructured":"Cai C, Wang Y (2020) A note on over-smoothing for graph neural networks. CoRR. arXiv:2006.13318"},{"key":"6426_CR2","unstructured":"Chen M, Wei Z, Huang Z, et\u00a0al (2020) Simple and deep graph convolutional networks. In: Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, Proceedings of Machine Learning Research, vol 119. PMLR. pp 1725\u20131735. http:\/\/proceedings.mlr.press\/v119\/chen20v.html"},{"key":"6426_CR3","unstructured":"Glorot X, Bengio Y (2010) Understanding the difficulty of training deep feedforward neural networks. In: Teh YW, Titterington DM (eds) Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, AISTATS 2010, Chia Laguna Resort, Sardinia, Italy, May 13-15, 2010, JMLR Proceedings, vol\u00a09. JMLR.org, pp 249\u2013256, http:\/\/proceedings.mlr.press\/v9\/glorot10a.html"},{"key":"6426_CR4","doi-asserted-by":"crossref","unstructured":"Han X, Zhao T, Liu Y, et\u00a0al (2023) Mlpinit: embarrassingly simple GNN training acceleration with MLP initialization. In: The eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023. OpenReview.net. https:\/\/openreview.net\/pdf?id=P8YIphWNEGO","DOI":"10.1109\/ICASSP49357.2023.10097203"},{"key":"6426_CR5","doi-asserted-by":"crossref","unstructured":"He K, Zhang X, Ren S, et\u00a0al (2015) Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. arXiv:1502.01852","DOI":"10.1109\/ICCV.2015.123"},{"key":"6426_CR6","unstructured":"Hu W, Fey M, Zitnik M, et\u00a0al (2020) Open graph benchmark: Datasets for machine learning on graphs. In: Larochelle H, Ranzato M, Hadsell R, et\u00a0al (eds) Advances in neural information processing systems 33: annual conference on neural information processing systems 2020, NeurIPS 2020, December 6-12, 2020, virtual"},{"key":"6426_CR7","unstructured":"Jaiswal A, Wang P, Chen T, et\u00a0al (2022) Old can be gold: better gradient flow can make vanilla-gcns great again. In: NeurIPS"},{"key":"6426_CR8","unstructured":"Kipf TN, Welling M (2017) Semi-supervised classification with graph convolutional networks. In: 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, conference track proceedings. OpenReview.net. https:\/\/openreview.net\/forum?id=SJU4ayYgl"},{"key":"6426_CR9","unstructured":"Klicpera J, Bojchevski A, G\u00fcnnemann S (2019) Predict then propagate: Graph neural networks meet personalized pagerank. In: 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019. OpenReview.net. https:\/\/openreview.net\/forum?id=H1gL-2A9Ym"},{"key":"6426_CR10","unstructured":"Li J, Song Y, Song X, et\u00a0al (2023) On the initialization of graph neural networks. In: Krause A, Brunskill E, Cho K, et\u00a0al (eds) International Conference on Machine Learning, ICML 2023, 23-29 July 2023, Honolulu, Hawaii, USA, Proceedings of Machine Learning Research, vol 202. PMLR, pp 19911\u201319931. https:\/\/proceedings.mlr.press\/v202\/li23y.html"},{"key":"6426_CR11","doi-asserted-by":"crossref","unstructured":"Li Q, Han Z, Wu X (2018) Deeper insights into graph convolutional networks for semi-supervised learning. In: McIlraith SA, Weinberger KQ (eds) Proceedings of the thirty-second AAAI conference on artificial intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence (EAAI-18), New Orleans, Louisiana, USA, February 2-7, 2018. AAAI Press, pp 3538\u20133545. https:\/\/www.aaai.org\/ocs\/index.php\/AAAI\/AAAI18\/paper\/view\/16098","DOI":"10.1609\/aaai.v32i1.11604"},{"key":"6426_CR12","unstructured":"van\u00a0der Maaten L, Hinton G (2008) Visualizing data using t-SNE. Journal of Machine Learning Research 9:2579\u20132605. http:\/\/www.jmlr.org\/papers\/v9\/vandermaaten08a.html"},{"key":"6426_CR13","unstructured":"Morris C, Kriege NM, Bause F, et\u00a0al (2020) Tudataset: a collection of benchmark datasets for learning with graphs. CoRR. arXiv:2007.08663"},{"key":"6426_CR14","unstructured":"Oono K, Suzuki T (2020) Graph neural networks exponentially lose expressive power for node classification. In: 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. https:\/\/openreview.net\/forum?id=S1ldO2EFPr"},{"key":"6426_CR15","unstructured":"Rong Y, Huang W, Xu T, et\u00a0al (2020) Dropedge: Towards deep graph convolutional networks on node classification. In: 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. https:\/\/openreview.net\/forum?id=Hkx1qkrKPr"},{"key":"6426_CR16","unstructured":"Shchur O, Mumme M, Bojchevski A, et\u00a0al (2018) Pitfalls of graph neural network evaluation. CoRR. arXiv:1811.05868"},{"key":"6426_CR17","doi-asserted-by":"crossref","unstructured":"Tao T, Vu V (2008) Random matrices: the circular law. 0708.2895","DOI":"10.1142\/S0219199708002788"},{"key":"6426_CR18","unstructured":"Velickovic P, Cucurull G, Casanova A, et\u00a0al (2018) Graph attention networks. In: 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net. https:\/\/openreview.net\/forum?id=rJXMpikCZ"},{"key":"6426_CR19","unstructured":"Wu Z, Ramsundar B, Feinberg EN, et\u00a0al (2017) Moleculenet: A benchmark for molecular machine learning. CoRR. arXiv:1703.00564"},{"key":"6426_CR20","unstructured":"Xu K, Li C, Tian Y, et\u00a0al (2018) Representation learning on graphs with jumping knowledge networks. In: Dy JG, Krause A (eds) Proceedings of the 35th International Conference on Machine Learning, ICML 2018, Stockholmsm\u00e4ssan, Stockholm, Sweden, July 10-15, 2018, Proceedings of Machine Learning Research, vol\u00a080. PMLR, pp 5449\u20135458. http:\/\/proceedings.mlr.press\/v80\/xu18c.html"},{"key":"6426_CR21","unstructured":"Zhao L, Akoglu L (2020) Pairnorm: tackling oversmoothing in gnns. In: 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. https:\/\/openreview.net\/forum?id=rkecl1rtwB"},{"key":"6426_CR22","unstructured":"Zhou K, Huang X, Zha D, et\u00a0al (2021) Dirichlet energy constrained learning for deep graph neural networks. In: Ranzato M, Beygelzimer A, Dauphin YN, et\u00a0al (eds) Advances in neural information processing systems 34: annual conference on Neural Information Processing Systems 2021, NeurIPS 2021, December 6-14, 2021, virtual, pp 21834\u201321846"}],"container-title":["Applied Intelligence"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10489-025-06426-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10489-025-06426-0\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10489-025-06426-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,19]],"date-time":"2025-09-19T19:36:42Z","timestamp":1758310602000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10489-025-06426-0"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4,7]]},"references-count":22,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2025,5]]}},"alternative-id":["6426"],"URL":"https:\/\/doi.org\/10.1007\/s10489-025-06426-0","relation":{},"ISSN":["0924-669X","1573-7497"],"issn-type":[{"value":"0924-669X","type":"print"},{"value":"1573-7497","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,4,7]]},"assertion":[{"value":"1 March 2025","order":1,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"7 April 2025","order":2,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no conflict of interest.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of Interest"}}],"article-number":"632"}}