Stage-wise Fine-tuning for Graph-to-Text Generation
arXiv:2105.08021 · doi:10.18653/v1/2021.acl-srw.2
Abstract
Graph-to-text generation has benefited from pre-trained language models (PLMs) in achieving better performance than structured graph encoders. However, they fail to fully utilize the structure information of the input graph. In this paper, we aim to further improve the performance of the pre-trained language model by proposing a structured graph-to-text model with a two-step fine-tuning mechanism which first fine-tunes the model on Wikipedia before adapting to the graph-to-text generation. In addition to using the traditional token and position embeddings to encode the knowledge graph (KG), we propose a novel tree-level embedding method to capture the inter-dependency structures of the input graph. This new approach has significantly improved the performance of all text generation metrics for the English WebNLG 2017 dataset.
arXiv comment10 pages, Accepted by Proceedings of ACL-IJCNLP 2021 Student Research Workshop, Code and Resources at https://github.com/EagleW/Stage-wise-Fine-tuning
References in corpus (6)
- BERTScore: Evaluating Text Generation with BERT
- TAPAS: Weakly Supervised Table Parsing via Pre-training
- TabFact: A Large-scale Dataset for Table-based Fact Verification
- Modeling Global and Local Node Contexts for Text Generation from Knowledge Graphs
- Describing a Knowledge Base
- DART: Open-Domain Structured Data Record to Text Generation