Matches in SemOpenAlex for { <https://semopenalex.org/work/W4320342772> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W4320342772 abstract "Self-supervised node representation learning aims to learn node representations from unlabelled graphs that rival the supervised counterparts. The key towards learning informative node representations lies in how to effectively gain contextual information from the graph structure. In this work, we present simple-yet-effective self-supervised node representation learning via aligning the hidden representations of nodes and their neighbourhood. Our first idea achieves such node-to-neighbourhood alignment by directly maximizing the mutual information between their representations, which, we prove theoretically, plays the role of graph smoothing. Our framework is optimized via a surrogate contrastive loss and a Topology-Aware Positive Sampling (TAPS) strategy is proposed to sample positives by considering the structural dependencies between nodes, which enables offline positive selection. Considering the excessive memory overheads of contrastive learning, we further propose a negative-free solution, where the main contribution is a Graph Signal Decorrelation (GSD) constraint to avoid representation collapse and over-smoothing. The GSD constraint unifies some of the existing constraints and can be used to derive new implementations to combat representation collapse. By applying our methods on top of simple MLP-based node representation encoders, we learn node representations that achieve promising node classification performance on a set of graph-structured datasets from small- to large-scale." @default.
- W4320342772 created "2023-02-13" @default.
- W4320342772 creator A5021350799 @default.
- W4320342772 creator A5046454314 @default.
- W4320342772 creator A5049078993 @default.
- W4320342772 date "2023-02-09" @default.
- W4320342772 modified "2023-10-16" @default.
- W4320342772 title "Self-Supervised Node Representation Learning via Node-to-Neighbourhood Alignment" @default.
- W4320342772 doi "https://doi.org/10.48550/arxiv.2302.04626" @default.
- W4320342772 hasPublicationYear "2023" @default.
- W4320342772 type Work @default.
- W4320342772 citedByCount "0" @default.
- W4320342772 crossrefType "posted-content" @default.
- W4320342772 hasAuthorship W4320342772A5021350799 @default.
- W4320342772 hasAuthorship W4320342772A5046454314 @default.
- W4320342772 hasAuthorship W4320342772A5049078993 @default.
- W4320342772 hasBestOaLocation W43203427721 @default.
- W4320342772 hasConcept C127413603 @default.
- W4320342772 hasConcept C132525143 @default.
- W4320342772 hasConcept C154945302 @default.
- W4320342772 hasConcept C31972630 @default.
- W4320342772 hasConcept C3770464 @default.
- W4320342772 hasConcept C41008148 @default.
- W4320342772 hasConcept C59404180 @default.
- W4320342772 hasConcept C62611344 @default.
- W4320342772 hasConcept C66938386 @default.
- W4320342772 hasConcept C80444323 @default.
- W4320342772 hasConceptScore W4320342772C127413603 @default.
- W4320342772 hasConceptScore W4320342772C132525143 @default.
- W4320342772 hasConceptScore W4320342772C154945302 @default.
- W4320342772 hasConceptScore W4320342772C31972630 @default.
- W4320342772 hasConceptScore W4320342772C3770464 @default.
- W4320342772 hasConceptScore W4320342772C41008148 @default.
- W4320342772 hasConceptScore W4320342772C59404180 @default.
- W4320342772 hasConceptScore W4320342772C62611344 @default.
- W4320342772 hasConceptScore W4320342772C66938386 @default.
- W4320342772 hasConceptScore W4320342772C80444323 @default.
- W4320342772 hasLocation W43203427721 @default.
- W4320342772 hasOpenAccess W4320342772 @default.
- W4320342772 hasPrimaryLocation W43203427721 @default.
- W4320342772 hasRelatedWork W2804057010 @default.
- W4320342772 hasRelatedWork W2964114465 @default.
- W4320342772 hasRelatedWork W3118281188 @default.
- W4320342772 hasRelatedWork W3133823961 @default.
- W4320342772 hasRelatedWork W3160144061 @default.
- W4320342772 hasRelatedWork W3168493052 @default.
- W4320342772 hasRelatedWork W4285218279 @default.
- W4320342772 hasRelatedWork W4287391817 @default.
- W4320342772 hasRelatedWork W4309038730 @default.
- W4320342772 hasRelatedWork W4309795810 @default.
- W4320342772 isParatext "false" @default.
- W4320342772 isRetracted "false" @default.
- W4320342772 workType "article" @default.