Modeling Global and Local Node Contexts for Text Generation from Knowledge Graphs

By Leonardo F. R. Ribeiro, Yue Zhang, Claire Gardent, Iryna Gurevych. 2020

82 citations recorded by the source corpus.

Abstract Recent graph-to-text models generate text from graph-based data using either global or local aggregation to learn node representations. Global node encoding allows explicit communication between two distant nodes, thereby neglecting graph topology as all nodes are directly connected. In contrast, local node encoding considers the relations between neighbor nodes capturing the graph structure, but it can fail to capture long-range relations. In this work, we gather both encoding strategies, proposing novel neural models that encode an input graph combining both global and local node contexts, in order to learn better contextualized node embeddings. In our experiments, we demonstrate that our approaches lead to significant improvements on two graph-to-text datasets achieving BLEU scores of 18.01 on the AGENDA dataset, and 63.69 on the WebNLG dataset for seen categories, outperforming state-of-the-art models by 3.7 and 3.1 points, respectively.1

Subfields: Artificial Intelligence, Natural Language Generation, Knowledge Graphs & Representation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Hinrich Schütze
2. Philip S. Yu
3. Iryna Gurevych
4. Leonardo F. R. Ribeiro
5. Lichao Sun
6. Martin Schmitt
7. Mohit Bansal
8. Zhiling Yan

In Natural Language Generation

1. Hinrich Schütze
2. Martin Schmitt
3. Leonardo F. R. Ribeiro
4. Iryna Gurevych

In Knowledge Graphs & Representation

1. Hinrich Schütze
2. Philip S. Yu
3. Iryna Gurevych
4. Torsten Zesch

Citations in the corpus

• Investigating Pretrained Language Models for Graph-to-Text Generation (2020, methodology): “While recent works have shown the benefit of explicitly encoding the graph structure in graph-to-text generation (Song et al., 2018; Ribeiro et al., 2019, 2020; Schmitt et al., 2020; Zhao et al., 2020a, to name a few), our approaches based on PLMs consistently outperform these models, even though…”
• A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT (2023, background): “[180] focus on utilizing local and global node encoding strategies jointly to capture complementary information from graph contexts.”