# Segmented Graph-Bert for Graph Instance Modeling

Source: [Segmented Graph-Bert for Graph Instance Modeling](<https://www.sota2.com/research/paper/68fa67051e209131cea13dd4>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T03:12:05.320Z

Authors: Jiawei Zhang

Published: 2020-02-09T04:55:07.000Z

## Abstract

In graph instance representation learning, both the diverse graph instance sizes and the graph node orderless property have been the major obstacles that render existing representation learning models fail to work. In this paper, we will examine the effectiveness of GRAPH-BERT on graph instance representation learning, which was designed for node representation learning tasks originally. To adapt GRAPH-BERT to the new problem settings, we re-design it with a segmented architecture instead, which is also named as SEG-BERT (Segmented GRAPH-BERT) for reference simplicity in this paper. SEG-BERT involves no node-order-variant inputs or functional components anymore, and it can handle the graph node orderless property naturally. What's more, SEG-BERT has a segmented architecture and introduces three different strategies to unify the graph instance sizes, i.e., full-input, padding/pruning and segment shifting, respectively. SEG-BERT is pre-trainable in an unsupervised manner, which can be further transferred to new tasks directly or with necessary fine-tuning. We have tested the effectiveness of SEG-BERT with experiments on seven graph instance benchmark datasets, and SEG-BERT can out-perform the comparison methods on six out of them with significant performance advantages.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Graph Classification on PTC](<https://www.sota2.com/research/sota/graph-classification-on-ptc.md>): Graph Classification; PTC; Accuracy: 68.86.

- [Graph Classification on IMDB-B](<https://www.sota2.com/research/sota/graph-classification-on-imdb-b.md>): Graph Classification; IMDB-B; Accuracy: 77.2.

- [Graph Classification on IMDB-M](<https://www.sota2.com/research/sota/graph-classification-on-imdb-m.md>): Graph Classification; IMDB-M; Accuracy: 53.4.

- [Graph Classification on MUTAG](<https://www.sota2.com/research/sota/graph-classification-on-mutag.md>): Graph Classification; MUTAG; Accuracy: 90.85.

- [Graph Classification on PROTEINS](<https://www.sota2.com/research/sota/graph-classification-on-proteins.md>): Graph Classification; PROTEINS; Accuracy: 77.09.

- [Graph Classification on COLLAB](<https://www.sota2.com/research/sota/graph-classification-on-collab.md>): Graph Classification; COLLAB; Accuracy: 78.42.

- [Graph Classification on NCI1](<https://www.sota2.com/research/sota/graph-classification-on-nci1.md>): Graph Classification; NCI1; Accuracy: 70.15.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2002.03283>)

[PDF](<https://arxiv.org/pdf/2002.03283.pdf>)

[Implementation](<https://github.com/jwzhanggy/SEG-BERT>)
