$\rm A^2Q$: Aggregation-Aware Quantization for Graph Neural Networks

TL;DR

This paper introduces A^2Q, a novel aggregation-aware mixed-precision quantization method for GNNs that automatically assigns bitwidths to nodes, significantly reducing model size and inference latency with minimal accuracy loss.

Contribution

The paper proposes A^2Q, which exploits GNN topology for adaptive quantization, including a local gradient training method and a nearest neighbor strategy for unseen graphs.

Findings

Achieves up to 18.6x compression with negligible accuracy loss.

Outperforms state-of-the-art quantization methods by up to 11.4% in accuracy.

Provides up to 2x speedup on dedicated hardware.

Abstract

As graph data size increases, the vast latency and memory consumption during inference pose a significant challenge to the real-world deployment of Graph Neural Networks (GNNs). While quantization is a powerful approach to reducing GNNs complexity, most previous works on GNNs quantization fail to exploit the unique characteristics of GNNs, suffering from severe accuracy degradation. Through an in-depth analysis of the topology of GNNs, we observe that the topology of the graph leads to significant differences between nodes, and most of the nodes in a graph appear to have a small aggregation value. Motivated by this, in this paper, we propose the Aggregation-Aware mixed-precision Quantization ($\rm A^2Q$) for GNNs, where an appropriate bitwidth is automatically learned and assigned to each node in the graph. To mitigate the vanishing gradient problem caused by sparse connections between nodes, we propose a Local Gradient method to serve the quantization error of the node features as the supervision during training. We also develop a Nearest Neighbor Strategy to deal with the generalization on unseen graphs. Extensive experiments on eight public node-level and graph-level datasets demonstrate the generality and robustness of our proposed method. Compared to the FP32 models, our method can achieve up to a 18.6x (i.e., 1.70bit) compression ratio with negligible accuracy degradation. Morever, compared to the state-of-the-art quantization method, our method can achieve up to 11.4\% and 9.5\% accuracy improvements on the node-level and graph-level tasks, respectively, and up to 2x speedup on a dedicated hardware accelerator.