Deep Graph Convolutional Encoders for Structured Data to Text Generation

Marcheggiani, Diego; Perez-Beltrachini, Laura

doi:10.18653/v1/w18-6501

Cited by 106 publications

(88 citation statements)

References 27 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Stacking layers was demonstrated to be effective in graph-to-sequence approaches (Marcheggiani and Perez Beltrachini, 2018;Koncel-Kedziorski et al, 2019;Damonte and Cohen, 2019) and allows us to test for their contributions to the system performance more easily. We employ different GNNs for both graph encoders (Section 3.3).…”

Section: Dual Graph Encodermentioning

confidence: 99%

Enhancing AMR-to-Text Generation with Dual Graph Representations

Ribeiro¹,

Gardent

Gurevych³

2019

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conferen

View full text Add to dashboard Cite

Generating text from graph-based data, such as Abstract Meaning Representation (AMR), is a challenging task due to the inherent difficulty in how to properly encode the structure of a graph with labeled edges. To address this difficulty, we propose a novel graph-to-sequence model that encodes different but complementary perspectives of the structural information contained in the AMR graph. The model learns parallel top-down and bottom-up representations of nodes capturing contrasting views of the graph. We also investigate the use of different node message passing strategies, employing different state-of-the-art graph encoders to compute node representations based on incoming and outgoing perspectives. In our experiments, we demonstrate that the dual graph representation leads to improvements in AMR-to-text generation, achieving state-ofthe-art results on two AMR datasets 1 .

show abstract

Section: Dual Graph Encodermentioning

confidence: 99%

Enhancing AMR-to-Text Generation with Dual Graph Representations

Ribeiro¹,

Gardent

Gurevych³

2019

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conferen

View full text Add to dashboard Cite

show abstract

“…Previous studies work on embedding-based model (Nguyen et al, 2018; and entity alignment model (Chen et al, 2017;Trisedya et al, 2019) to enrich a knowledge base. Following the success of the sequence-to-sequence architecture (Bahdanau et al, 2015) for generating sentences from structured data (Marcheggiani and Perez-Beltrachini, 2018;Trisedya et al, 2018), we employ this architecture to do the opposite, which is extracting triples from a sentence.…”

Section: Introductionmentioning

confidence: 99%

Neural Relation Extraction for Knowledge Base Enrichment

Trisedya¹,

Weikum

Qi³

et al. 2019

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

View full text Add to dashboard Cite

We study relation extraction for knowledge base (KB) enrichment. Specifically, we aim to extract entities and their relationships from sentences in the form of triples and map the elements of the extracted triples to an existing KB in an end-to-end manner. Previous studies focus on the extraction itself and rely on Named Entity Disambiguation (NED) to map triples into the KB space. This way, NED errors may cause extraction errors that affect the overall precision and recall. To address this problem, we propose an end-to-end relation extraction model for KB enrichment based on a neural encoder-decoder model. We collect high-quality training data by distant supervision with co-reference resolution and paraphrase detection. We propose an n-gram based attention model that captures multi-word entity names in a sentence. Our model employs jointly learned word and entity embeddings to support named entity disambiguation. Finally, our model uses a modified beam search and a triple classifier to help generate high-quality triples. Our model outperforms state-of-theart baselines by 15.51% and 8.38% in terms of F1 score on two real-world datasets.

show abstract

“…Another critical open question in our framework is whether the surface generator will be able to generate surfaces representative enough to allow for generalization to real examples. Current NLG systems are increasingly capable of structured text generation (Marcheggiani and Perez-Beltrachini, 2018), and though they produce relatively short surfaces, we believe that coupling them with the generated action graphs is a promising approach to scaling up to longer sequences while maintaining coherence. Such systems can use sentence-level semantic parses as training data, meaning they can leverage existing weakly-supervised shallow parsing techniques.…”

Section: Discussionmentioning

confidence: 99%

Playing by the Book: An Interactive Game Approach for Action Graph Extraction from Text

Tamari¹,

Shindo²,

Shahaf³

et al. 2019

Proceedings of the Workshop on Extracting Structured Knowledge From Scientific Publications

View full text Add to dashboard Cite

Understanding procedural text requires tracking entities, actions and effects as the narrative unfolds. We focus on the challenging realworld problem of action-graph extraction from materials science papers, where language is highly specialized and data annotation is expensive and scarce. We propose a novel approach, TEXT2QUEST, where procedural text is interpreted as instructions for an interactive game. A learning agent completes the game by executing the procedure correctly in a textbased simulated lab environment. The framework can complement existing approaches and enables richer forms of learning compared to static texts. We discuss potential limitations and advantages of the approach, and release a prototype proof-of-concept, hoping to encourage research in this direction.

show abstract

Deep Graph Convolutional Encoders for Structured Data to Text Generation

Cited by 106 publications

References 27 publications

Enhancing AMR-to-Text Generation with Dual Graph Representations

Enhancing AMR-to-Text Generation with Dual Graph Representations

Neural Relation Extraction for Knowledge Base Enrichment

Playing by the Book: An Interactive Game Approach for Action Graph Extraction from Text

Contact Info

Product

Resources

About