CycleGT: Unsupervised Graph-to-Text and Text-to-Graph Generation via Cycle Training

TL;DR

CycleGT introduces an unsupervised cycle training approach for graph-to-text and text-to-graph conversion, effectively addressing data scarcity by leveraging non-parallel data and achieving competitive performance.

Contribution

It proposes a novel cycle training framework for unsupervised G2T and T2G tasks, enabling learning from non-parallel data without requiring large annotated datasets.

Findings

Achieves comparable performance to supervised models on WebNLG datasets.

Outperforms other unsupervised baselines on GenWiki dataset.

Demonstrates effectiveness in overcoming data scarcity in G2T and T2G tasks.

Abstract

Two important tasks at the intersection of knowledge graphs and natural language processing are graph-to-text (G2T) and text-to-graph (T2G) conversion. Due to the difficulty and high cost of data collection, the supervised data available in the two fields are usually on the magnitude of tens of thousands, for example, 18K in the WebNLG~2017 dataset after preprocessing, which is far fewer than the millions of data for other tasks such as machine translation. Consequently, deep learning models for G2T and T2G suffer largely from scarce training data. We present CycleGT, an unsupervised training method that can bootstrap from fully non-parallel graph and text data, and iteratively back translate between the two forms. Experiments on WebNLG datasets show that our unsupervised model trained on the same number of data achieves performance on par with several fully supervised models. Further experiments on the non-parallel GenWiki dataset verify that our method performs the best among unsupervised baselines. This validates our framework as an effective approach to overcome the data scarcity problem in the fields of G2T and T2G. Our code is available at https://github.com/QipengGuo/CycleGT.