Matches in SemOpenAlex for { <https://semopenalex.org/work/W4382202907> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4382202907 endingPage "10360" @default.
- W4382202907 startingPage "10351" @default.
- W4382202907 abstract "Recent years have witnessed the great success of Graph Neural Networks (GNNs) in handling graph-related tasks. However, MLPs remain the primary workhorse for practical industrial applications due to their desirable inference efficiency and scalability. To reduce their gaps, one can directly distill knowledge from a well-designed teacher GNN to a student MLP, which is termed as GNN-to-MLP distillation. However, the process of distillation usually entails a loss of information, and ``which knowledge patterns of GNNs are more likely to be left and distilled into MLPs? becomes an important question. In this paper, we first factorize the knowledge learned by GNNs into low- and high-frequency components in the spectral domain and then derive their correspondence in the spatial domain. Furthermore, we identified a potential information drowning problem for existing GNN-to-MLP distillation, i.e., the high-frequency knowledge of the pre-trained GNNs may be overwhelmed by the low-frequency knowledge during distillation; we have described in detail what it represents, how it arises, what impact it has, and how to deal with it. In this paper, we propose an efficient Full-Frequency GNN-to-MLP (FF-G2M) distillation framework, which extracts both low-frequency and high-frequency knowledge from GNNs and injects it into MLPs. Extensive experiments show that FF-G2M improves over the vanilla MLPs by 12.6% and outperforms its corresponding teacher GNNs by 2.6% averaged over six graph datasets and three common GNN architectures." @default.
- W4382202907 created "2023-06-28" @default.
- W4382202907 creator A5020499056 @default.
- W4382202907 creator A5047726315 @default.
- W4382202907 creator A5066706130 @default.
- W4382202907 creator A5068887242 @default.
- W4382202907 creator A5082786719 @default.
- W4382202907 date "2023-06-26" @default.
- W4382202907 modified "2023-09-30" @default.
- W4382202907 title "Extracting Low-/High- Frequency Knowledge from Graph Neural Networks and Injecting It into MLPs: An Effective GNN-to-MLP Distillation Framework" @default.
- W4382202907 doi "https://doi.org/10.1609/aaai.v37i9.26232" @default.
- W4382202907 hasPublicationYear "2023" @default.
- W4382202907 type Work @default.
- W4382202907 citedByCount "1" @default.
- W4382202907 countsByYear W43822029072023 @default.
- W4382202907 crossrefType "journal-article" @default.
- W4382202907 hasAuthorship W4382202907A5020499056 @default.
- W4382202907 hasAuthorship W4382202907A5047726315 @default.
- W4382202907 hasAuthorship W4382202907A5066706130 @default.
- W4382202907 hasAuthorship W4382202907A5068887242 @default.
- W4382202907 hasAuthorship W4382202907A5082786719 @default.
- W4382202907 hasBestOaLocation W43822029071 @default.
- W4382202907 hasConcept C111919701 @default.
- W4382202907 hasConcept C119857082 @default.
- W4382202907 hasConcept C124101348 @default.
- W4382202907 hasConcept C132525143 @default.
- W4382202907 hasConcept C153180895 @default.
- W4382202907 hasConcept C154945302 @default.
- W4382202907 hasConcept C185592680 @default.
- W4382202907 hasConcept C19118579 @default.
- W4382202907 hasConcept C204030448 @default.
- W4382202907 hasConcept C207685749 @default.
- W4382202907 hasConcept C2776214188 @default.
- W4382202907 hasConcept C31972630 @default.
- W4382202907 hasConcept C41008148 @default.
- W4382202907 hasConcept C43617362 @default.
- W4382202907 hasConcept C48044578 @default.
- W4382202907 hasConcept C77088390 @default.
- W4382202907 hasConcept C80444323 @default.
- W4382202907 hasConcept C98045186 @default.
- W4382202907 hasConceptScore W4382202907C111919701 @default.
- W4382202907 hasConceptScore W4382202907C119857082 @default.
- W4382202907 hasConceptScore W4382202907C124101348 @default.
- W4382202907 hasConceptScore W4382202907C132525143 @default.
- W4382202907 hasConceptScore W4382202907C153180895 @default.
- W4382202907 hasConceptScore W4382202907C154945302 @default.
- W4382202907 hasConceptScore W4382202907C185592680 @default.
- W4382202907 hasConceptScore W4382202907C19118579 @default.
- W4382202907 hasConceptScore W4382202907C204030448 @default.
- W4382202907 hasConceptScore W4382202907C207685749 @default.
- W4382202907 hasConceptScore W4382202907C2776214188 @default.
- W4382202907 hasConceptScore W4382202907C31972630 @default.
- W4382202907 hasConceptScore W4382202907C41008148 @default.
- W4382202907 hasConceptScore W4382202907C43617362 @default.
- W4382202907 hasConceptScore W4382202907C48044578 @default.
- W4382202907 hasConceptScore W4382202907C77088390 @default.
- W4382202907 hasConceptScore W4382202907C80444323 @default.
- W4382202907 hasConceptScore W4382202907C98045186 @default.
- W4382202907 hasIssue "9" @default.
- W4382202907 hasLocation W43822029071 @default.
- W4382202907 hasLocation W43822029072 @default.
- W4382202907 hasOpenAccess W4382202907 @default.
- W4382202907 hasPrimaryLocation W43822029071 @default.
- W4382202907 hasRelatedWork W1525643724 @default.
- W4382202907 hasRelatedWork W2067938758 @default.
- W4382202907 hasRelatedWork W2163489736 @default.
- W4382202907 hasRelatedWork W2302028273 @default.
- W4382202907 hasRelatedWork W2333420780 @default.
- W4382202907 hasRelatedWork W2364921833 @default.
- W4382202907 hasRelatedWork W2382623646 @default.
- W4382202907 hasRelatedWork W2961085424 @default.
- W4382202907 hasRelatedWork W3087771547 @default.
- W4382202907 hasRelatedWork W4386590857 @default.
- W4382202907 hasVolume "37" @default.
- W4382202907 isParatext "false" @default.
- W4382202907 isRetracted "false" @default.
- W4382202907 workType "article" @default.