NP$^2$L: Negative Pseudo Partial Labels Extraction for Graph Neural Networks

TL;DR

This paper introduces NP$^2$L, a method for extracting negative pseudo partial labels to improve graph neural network learning, achieving state-of-the-art results in link prediction and node classification.

Contribution

The paper proposes a novel negative pseudo partial labels extraction method that enhances GNN training by constructing signed graphs with more accurate pseudo labels.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Effectively improves GNN learning through negative pseudo partial labels.

Demonstrates the importance of label selection in pseudo label utilization.

Abstract

How to utilize the pseudo labels has always been a research hotspot in machine learning. However, most methods use pseudo labels as supervised training, and lack of valid assessing for their accuracy. Moreover, applications of pseudo labels in graph neural networks (GNNs) oversee the difference between graph learning and other machine learning tasks such as message passing mechanism. Aiming to address the first issue, we found through a large number of experiments that the pseudo labels are more accurate if they are selected by not overlapping partial labels and defined as negative node pairs relations. Therefore, considering the extraction based on pseudo and partial labels, negative edges are constructed between two nodes by the negative pseudo partial labels extraction (NP$^2$E) module. With that, a signed graph are built containing highly accurate pseudo labels information from the original graph, which effectively assists GNN in learning at the message-passing level, provide one solution to the second issue. Empirical results about link prediction and node classification tasks on several benchmark datasets demonstrate the effectiveness of our method. State-of-the-art performance is achieved on the both tasks.