DepWiGNN: A Depth-wise Graph Neural Network for Multi-hop Spatial Reasoning in Text

TL;DR

DepWiGNN introduces a depth-wise graph neural network that effectively captures long-range spatial dependencies in text-based reasoning tasks, overcoming over-smoothing issues of traditional GNNs.

Contribution

The paper proposes a novel depth-wise GNN with a node memory scheme, enabling multi-hop spatial reasoning without deep stacking of layers.

Findings

DepWiGNN outperforms existing methods on two spatial reasoning datasets.

It demonstrates superior ability to capture long dependencies in graphs.

The approach mitigates over-smoothing in GNNs for multi-hop reasoning.

Abstract

Spatial reasoning in text plays a crucial role in various real-world applications. Existing approaches for spatial reasoning typically infer spatial relations from pure text, which overlooks the gap between natural language and symbolic structures. Graph neural networks (GNNs) have showcased exceptional proficiency in inducing and aggregating symbolic structures. However, classical GNNs face challenges in handling multi-hop spatial reasoning due to the over-smoothing issue, i.e., the performance decreases substantially as the number of graph layers increases. To cope with these challenges, we propose a novel Depth-Wise Graph Neural Network (DepWiGNN). Specifically, we design a novel node memory scheme and aggregate the information over the depth dimension instead of the breadth dimension of the graph, which empowers the ability to collect long dependencies without stacking multiple layers. Experimental results on two challenging multi-hop spatial reasoning datasets show that DepWiGNN outperforms existing spatial reasoning methods. The comparisons with the other three GNNs further demonstrate its superiority in capturing long dependency in the graph.