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

Abstract

Graph Neural Networks (GNNs) are popular for graph machine learning and have shown great results on wide node classification tasks. Yet, they are less popular for practical deployments in the industry owing to their scalability challenges incurred by data dependency. Namely, GNN inference depends on neighbor nodes multiple hops away from the target, and fetching them burdens latency-constrained applications. Existing inference acceleration methods like pruning and quantization can speed up GNNs by reducing Multiplication-and-ACcumulation (MAC) operations, but the improvements are limited given the data dependency is not resolved. Conversely, multi-layer perceptrons (MLPs) have no graph dependency and infer much faster than GNNs, even though they are less accurate than GNNs for node classification in general. Motivated by these complementary strengths and weaknesses, we bring GNNs and MLPs together via knowledge distillation (KD). Our work shows that the performance of MLPs can be improved by large margins with GNN KD. We call the distilled MLPs Graph-less Neural Networks (GLNNs) as they have no inference graph dependency. We show that GLNNs with competitive accuracy infer faster than GNNs by 146X-273X and faster than other acceleration methods by 14X-27X. Under a production setting involving both transductive and inductive predictions across 7 datasets, GLNN accuracies improve over stand-alone MLPs by 12.36% on average and match GNNs on 6/7 datasets. Comprehensive analysis shows when and why GLNNs can achieve competitive accuracies to GNNs and suggests GLNN as a handy choice for latency-constrained applications.

References in corpus (11)

• Distilling the Knowledge in a Neural Network
• Semi-Supervised Classification with Graph Convolutional Networks
• MLP-Mixer: An all-MLP Architecture for Vision
• Simplifying Graph Convolutional Networks
• Deep Learning with Limited Numerical Precision
• FastGCN: Fast Learning with Graph Convolutional Networks via Importance Sampling
• Layer-Dependent Importance Sampling for Training Deep and Large Graph Convolutional Networks
• Do You Even Need Attention? A Stack of Feed-Forward Layers Does Surprisingly Well on ImageNet
• Learned Low Precision Graph Neural Networks
• On Graph Neural Networks versus Graph-Augmented MLPs
• Extract the Knowledge of Graph Neural Networks and Go Beyond it: An Effective Knowledge Distillation Framework

Cited by in corpus (4)

• On Representation Knowledge Distillation for Graph Neural Networks
• Graph-less Collaborative Filtering
• LightGNN: Simple Graph Neural Network for Recommendation
• Cold Brew: Distilling Graph Node Representations with Incomplete or Missing Neighborhoods