Matches in SemOpenAlex for { <https://semopenalex.org/work/W4298054246> ?p ?o ?g. }
Showing items 1 to 59 of
59
with 100 items per page.
- W4298054246 abstract "Contrastive learning has recently achieved remarkable success in many domains including graphs. However contrastive loss, especially for graphs, requires a large number of negative samples which is unscalable and computationally prohibitive with a quadratic time complexity. Sub-sampling is not optimal and incorrect negative sampling leads to sampling bias. In this work, we propose a meta-node based approximation technique that can (a) proxy all negative combinations (b) in quadratic cluster size time complexity, (c) at graph level, not node level, and (d) exploit graph sparsity. By replacing node-pairs with additive cluster-pairs, we compute the negatives in cluster-time at graph level. The resulting Proxy approximated meta-node Contrastive (PamC) loss, based on simple optimized GPU operations, captures the full set of negatives, yet is efficient with a linear time complexity. By avoiding sampling, we effectively eliminate sample bias. We meet the criterion for larger number of samples, thus achieving block-contrastiveness, which is proven to outperform pair-wise losses. We use learnt soft cluster assignments for the meta-node constriction, and avoid possible heterophily and noise added during edge creation. Theoretically, we show that real world graphs easily satisfy conditions necessary for our approximation. Empirically, we show promising accuracy gains over state-of-the-art graph clustering on 6 benchmarks. Importantly, we gain substantially in efficiency; up to 3x in training time, 1.8x in inference time and over 5x in GPU memory reduction." @default.
- W4298054246 created "2022-10-01" @default.
- W4298054246 creator A5040887350 @default.
- W4298054246 creator A5078468070 @default.
- W4298054246 creator A5083545460 @default.
- W4298054246 date "2022-09-28" @default.
- W4298054246 modified "2023-10-16" @default.
- W4298054246 title "Efficient block contrastive learning via parameter-free meta-node approximation" @default.
- W4298054246 doi "https://doi.org/10.48550/arxiv.2209.14067" @default.
- W4298054246 hasPublicationYear "2022" @default.
- W4298054246 type Work @default.
- W4298054246 citedByCount "0" @default.
- W4298054246 crossrefType "posted-content" @default.
- W4298054246 hasAuthorship W4298054246A5040887350 @default.
- W4298054246 hasAuthorship W4298054246A5078468070 @default.
- W4298054246 hasAuthorship W4298054246A5083545460 @default.
- W4298054246 hasBestOaLocation W42980542461 @default.
- W4298054246 hasConcept C11413529 @default.
- W4298054246 hasConcept C129844170 @default.
- W4298054246 hasConcept C132525143 @default.
- W4298054246 hasConcept C154945302 @default.
- W4298054246 hasConcept C165696696 @default.
- W4298054246 hasConcept C179799912 @default.
- W4298054246 hasConcept C2524010 @default.
- W4298054246 hasConcept C2776214188 @default.
- W4298054246 hasConcept C311688 @default.
- W4298054246 hasConcept C33923547 @default.
- W4298054246 hasConcept C38652104 @default.
- W4298054246 hasConcept C41008148 @default.
- W4298054246 hasConcept C80444323 @default.
- W4298054246 hasConceptScore W4298054246C11413529 @default.
- W4298054246 hasConceptScore W4298054246C129844170 @default.
- W4298054246 hasConceptScore W4298054246C132525143 @default.
- W4298054246 hasConceptScore W4298054246C154945302 @default.
- W4298054246 hasConceptScore W4298054246C165696696 @default.
- W4298054246 hasConceptScore W4298054246C179799912 @default.
- W4298054246 hasConceptScore W4298054246C2524010 @default.
- W4298054246 hasConceptScore W4298054246C2776214188 @default.
- W4298054246 hasConceptScore W4298054246C311688 @default.
- W4298054246 hasConceptScore W4298054246C33923547 @default.
- W4298054246 hasConceptScore W4298054246C38652104 @default.
- W4298054246 hasConceptScore W4298054246C41008148 @default.
- W4298054246 hasConceptScore W4298054246C80444323 @default.
- W4298054246 hasLocation W42980542461 @default.
- W4298054246 hasOpenAccess W4298054246 @default.
- W4298054246 hasPrimaryLocation W42980542461 @default.
- W4298054246 hasRelatedWork W1507793640 @default.
- W4298054246 hasRelatedWork W1543400885 @default.
- W4298054246 hasRelatedWork W2107287896 @default.
- W4298054246 hasRelatedWork W2156272239 @default.
- W4298054246 hasRelatedWork W2271929959 @default.
- W4298054246 hasRelatedWork W2355108509 @default.
- W4298054246 hasRelatedWork W2950696215 @default.
- W4298054246 hasRelatedWork W4301599260 @default.
- W4298054246 hasRelatedWork W4327520943 @default.
- W4298054246 hasRelatedWork W2179712692 @default.
- W4298054246 isParatext "false" @default.
- W4298054246 isRetracted "false" @default.
- W4298054246 workType "article" @default.