Matches in SemOpenAlex for { <https://semopenalex.org/work/W3193367109> ?p ?o ?g. }
- W3193367109 abstract "Self-supervised Learning (SSL) aims at learning representations of objects without relying on manual labeling. Recently, a number of SSL methods for graph representation learning have achieved performance comparable to SOTA semi-supervised GNNs. A Siamese network, which relies on data augmentation, is the popular architecture used in these methods. However, these methods rely on heuristically crafted data augmentation techniques. Furthermore, they use either contrastive terms or other tricks (e.g., asymmetry) to avoid trivial solutions that can occur in Siamese networks. In this study, we propose, GraphSurgeon, a novel SSL method for GNNs with the following features. First, instead of heuristics we propose a learnable data augmentation method that is jointly learned with the embeddings by leveraging the inherent signal encoded in the graph. In addition, we take advantage of the flexibility of the learnable data augmentation and introduce a new strategy that augments in the embedding space, called post augmentation. This strategy has a significantly lower memory overhead and run-time cost. Second, as it is difficult to sample truly contrastive terms, we avoid explicit negative sampling. Third, instead of relying on engineering tricks, we use a scalable constrained optimization objective motivated by Laplacian Eigenmaps to avoid trivial solutions. To validate the practical use of GraphSurgeon, we perform empirical evaluation using 14 public datasets across a number of domains and ranging from small to large scale graphs with hundreds of millions of edges. Our finding shows that GraphSurgeon is comparable to six SOTA semi-supervised and on par with five SOTA self-supervised baselines in node classification tasks. The source code is available at this https URL." @default.
- W3193367109 created "2021-08-30" @default.
- W3193367109 creator A5036955634 @default.
- W3193367109 creator A5041411651 @default.
- W3193367109 creator A5088533003 @default.
- W3193367109 date "2021-08-23" @default.
- W3193367109 modified "2023-10-01" @default.
- W3193367109 title "Jointly Learnable Data Augmentations for Self-Supervised GNNs" @default.
- W3193367109 cites W2097308346 @default.
- W3193367109 cites W2171590421 @default.
- W3193367109 cites W2519887557 @default.
- W3193367109 cites W2624431344 @default.
- W3193367109 cites W2766453196 @default.
- W3193367109 cites W2786915849 @default.
- W3193367109 cites W2804057010 @default.
- W3193367109 cites W2893944917 @default.
- W3193367109 cites W2900470550 @default.
- W3193367109 cites W2916106175 @default.
- W3193367109 cites W2952254971 @default.
- W3193367109 cites W2961295589 @default.
- W3193367109 cites W2975966238 @default.
- W3193367109 cites W2987283559 @default.
- W3193367109 cites W2994536315 @default.
- W3193367109 cites W2995914187 @default.
- W3193367109 cites W2998496395 @default.
- W3193367109 cites W3005644236 @default.
- W3193367109 cites W3033039844 @default.
- W3193367109 cites W3034978746 @default.
- W3193367109 cites W3035659929 @default.
- W3193367109 cites W3081963674 @default.
- W3193367109 cites W3087257704 @default.
- W3193367109 cites W3095121901 @default.
- W3193367109 cites W3101821705 @default.
- W3193367109 cites W3102419180 @default.
- W3193367109 cites W3107668149 @default.
- W3193367109 cites W3132402977 @default.
- W3193367109 cites W3146165332 @default.
- W3193367109 cites W3157286395 @default.
- W3193367109 cites W3159481202 @default.
- W3193367109 cites W3163602117 @default.
- W3193367109 cites W3167553825 @default.
- W3193367109 cites W3168925038 @default.
- W3193367109 cites W3169717035 @default.
- W3193367109 cites W3172495685 @default.
- W3193367109 cites W3172710079 @default.
- W3193367109 hasPublicationYear "2021" @default.
- W3193367109 type Work @default.
- W3193367109 sameAs 3193367109 @default.
- W3193367109 citedByCount "0" @default.
- W3193367109 crossrefType "posted-content" @default.
- W3193367109 hasAuthorship W3193367109A5036955634 @default.
- W3193367109 hasAuthorship W3193367109A5041411651 @default.
- W3193367109 hasAuthorship W3193367109A5088533003 @default.
- W3193367109 hasConcept C111919701 @default.
- W3193367109 hasConcept C119857082 @default.
- W3193367109 hasConcept C127705205 @default.
- W3193367109 hasConcept C132525143 @default.
- W3193367109 hasConcept C154945302 @default.
- W3193367109 hasConcept C184898388 @default.
- W3193367109 hasConcept C41008148 @default.
- W3193367109 hasConcept C41608201 @default.
- W3193367109 hasConcept C48044578 @default.
- W3193367109 hasConcept C77088390 @default.
- W3193367109 hasConcept C80444323 @default.
- W3193367109 hasConceptScore W3193367109C111919701 @default.
- W3193367109 hasConceptScore W3193367109C119857082 @default.
- W3193367109 hasConceptScore W3193367109C127705205 @default.
- W3193367109 hasConceptScore W3193367109C132525143 @default.
- W3193367109 hasConceptScore W3193367109C154945302 @default.
- W3193367109 hasConceptScore W3193367109C184898388 @default.
- W3193367109 hasConceptScore W3193367109C41008148 @default.
- W3193367109 hasConceptScore W3193367109C41608201 @default.
- W3193367109 hasConceptScore W3193367109C48044578 @default.
- W3193367109 hasConceptScore W3193367109C77088390 @default.
- W3193367109 hasConceptScore W3193367109C80444323 @default.
- W3193367109 hasLocation W31933671091 @default.
- W3193367109 hasOpenAccess W3193367109 @default.
- W3193367109 hasPrimaryLocation W31933671091 @default.
- W3193367109 hasRelatedWork W167670643 @default.
- W3193367109 hasRelatedWork W2151575489 @default.
- W3193367109 hasRelatedWork W2499981259 @default.
- W3193367109 hasRelatedWork W2742267824 @default.
- W3193367109 hasRelatedWork W2798908575 @default.
- W3193367109 hasRelatedWork W2809251119 @default.
- W3193367109 hasRelatedWork W2999162813 @default.
- W3193367109 hasRelatedWork W3007962981 @default.
- W3193367109 hasRelatedWork W3037584823 @default.
- W3193367109 hasRelatedWork W3099726174 @default.
- W3193367109 hasRelatedWork W3102419180 @default.
- W3193367109 hasRelatedWork W3121107455 @default.
- W3193367109 hasRelatedWork W3134028393 @default.
- W3193367109 hasRelatedWork W3167553825 @default.
- W3193367109 hasRelatedWork W3168925038 @default.
- W3193367109 hasRelatedWork W3185459701 @default.
- W3193367109 hasRelatedWork W3193376721 @default.
- W3193367109 hasRelatedWork W3209279086 @default.
- W3193367109 hasRelatedWork W3211996088 @default.
- W3193367109 hasRelatedWork W2185806431 @default.
- W3193367109 isParatext "false" @default.
- W3193367109 isRetracted "false" @default.