Bridging the Structural Gap Between Encoding and Decoding for Data-To-Text Generation

By Chao Zhao, Marilyn A. Walker, Snigdha Chaturvedi. 2020

103 citations recorded by the source corpus.

Generating sequential natural language descriptions from graph-structured data (e.g., knowledge graph) is challenging, partly because of the structural differences between the input graph and the output text. Hence, popular sequence-to-sequence models, which require serialized input, are not a natural fit for this task. Graph neural networks, on the other hand, can better encode the input graph but broaden the structural gap between the encoder and decoder, making faithful generation difficult. To narrow this gap, we propose DualEnc, a dual encoding model that can not only incorporate the graph structure, but can also cater to the linear structure of the output text. Empirical comparisons with strong single-encoder baselines demonstrate that dual encoding can significantly improve the quality of the generated text.

Subfields: Artificial Intelligence, Natural Language Generation

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

In Natural Language Generation

Citations in the corpus

Text-to-Text Pre-Training for Data-to-Text Tasks (2020, neutral): “• PlanEnc (Zhao et al., 2020), the current stateof-the-art system.”
Denoising Pre-Training and Data Augmentation Strategies for Enhanced RDF Verbalization with Transformers (2020, unknown): “Following this conclusion, several papers proposed graph-based encoder as a solution (Gao et al., 2020; Zhao et al., 2020; Moussallem et al., 2020).”
Investigating Pretrained Language Models for Graph-to-Text Generation (2020, methodology): “Recent approaches (Marcheggiani and Perez Bel-trachini, 2018; Song et al., 2018; Beck et al., 2018; Damonte and Cohen, 2019; Ribeiro et al., 2019; Zhao et al., 2020a; Schmitt et al., 2021; Ribeiro et al., 2021b) propose architectures based on GNNs to directly encode the graph structure, whereas…”
Improving Text-to-Text Pre-trained Models for the Graph-to-Text Task (2020, background): “To address this problem, triple-speciﬁc or graph-based encoders have been proposed to better process the KG triple input (Vougiouklis et al., 2018; Distiawan et al., 2018; Marcheggiani and Perez-Beltrachini, 2018; Schmitt et al., 2020; Zhao et al., 2020).”
A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT (2023, background): “DUALENC [175]: a KG-to-text generation model that bridges the structural gap between KG and”
ESPRIT: Explaining Solutions to Physical Reasoning Tasks (2020, background): “, 2016) have shown promising results on table-to-text tasks (Wiseman et al., 2017; Gehrmann et al., 2018; Puduppully et al., 2019a,b; Iso et al., 2019; Castro Ferreira et al., 2019; Zhao et al., 2020; Chen et al., 2020a).”
A Survey of Controllable Text Generation Using Transformer-based Pre-trained Language Models (2022, background): “For instance, the lexical-level overlap between the reference text containing the control elements and the generated text can be used to reflect how well the control conditions are met in the task of table/graph to text [102, 111, 112, 128, 131, 165] and story generation [33, 40].”
AggGen: Ordering and Aggregating while Generating (2021, methodology): “…works for comparison: (1) best-performing models reported on the WebNLG 2017 (seen) and E2E dataset, i.e. T5 (Kale and Rastogi, 2020), PlanEnc (Zhao et al., 2020), ADAPT (Gardent et al., 2017b), and TGen (Dušek and Jurčı́ček, 2016); (2) models with explicit planning, i.e. TILB-PIPE…”