English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Graph Neural Networks (GNNs) are popular for graph machine learning and haveshown great results on wide node classification tasks. Yet, they are lesspopular for practical deployments in the industry owing to their scalabilitychallenges incurred by data dependency. Namely, GNN inference depends onneighbor nodes multiple hops away from the target, and fetching them burdenslatency-constrained applications. Existing inference acceleration methods likepruning and quantization can speed up GNNs by reducingMultiplication-and-ACcumulation (MAC) operations, but the improvements arelimited given the data dependency is not resolved. Conversely, multi-layerperceptrons (MLPs) have no graph dependency and infer much faster than GNNs,even though they are less accurate than GNNs for node classification ingeneral. Motivated by these complementary strengths and weaknesses, we bringGNNs and MLPs together via knowledge distillation (KD). Our work shows that theperformance of MLPs can be improved by large margins with GNN KD. We call thedistilled MLPs Graph-less Neural Networks (GLNNs) as they have no inferencegraph dependency. We show that GLNNs with competitive accuracy infer fasterthan GNNs by 146X-273X and faster than other acceleration methods by 14X-27X.Under a production setting involving both transductive and inductivepredictions across 7 datasets, GLNN accuracies improve over stand-alone MLPs by12.36% on average and match GNNs on 6/7 datasets. Comprehensive analysis showswhen and why GLNNs can achieve competitive accuracies to GNNs and suggests GLNNas a handy choice for latency-constrained applications.

Graph-less Neural Networks: Teaching Old MLPs New Tricks via Distillation