Unsupervised Dual Paraphrasing for Two-stage Semantic Parsing

Cao, Ruisheng; Zhu, Su; Yang, Chenyu; Liu, Chen; Ma, Rao; Zhao, Yuchen; Chen, Lu; Yu, Kai

doi:10.18653/v1/2020.acl-main.608

Cited by 26 publications

(21 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For many Natural Language Processing (NLP) tasks, there exist many primal and dual tasks, such as open information narration (OIN) and open information extraction (OIE) (Sun et al, 2018), natural language understanding (NLU) and natural language generation (NLG) , semantic parsing and natural language generation (Ye et al, 2019;Cao et al, 2019Cao et al, , 2020, link prediction and entailment graph induction (Cao et al, 2019), query-to-response and response-to-query generation (Shen and Feng, 2020) and so on. The duality between the primal task and the dual task is considered as a constraint that both problems must share the same joint probability mutually.…”

Section: B1 Dual Learningmentioning

confidence: 99%

LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification

Zuo¹,

Cao²,

Chen³

et al. 2021

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Confer

View full text Add to dashboard Cite

Modern models for event causality identification (ECI) are mainly based on supervised learning, which are prone to the data lacking problem. Unfortunately, the existing NLPrelated augmentation methods cannot directly produce available data required for this task. To solve the data lacking problem, we introduce a new approach to augment training data for event causality identification, by iteratively generating new examples and classifying event causality in a dual learning framework. On the one hand, our approach is knowledge guided, which can leverage existing knowledge bases to generate well-formed new sentences. On the other hand, our approach employs a dual mechanism, which is a learnable augmentation framework, and can interactively adjust the generation process to generate task-related sentences. Experimental results on two benchmarks EventStoryLine and Causal-TimeBank show that 1) our method can augment suitable task-related training data for ECI; 2) our method outperforms previous methods on EventStoryLine and Causal-TimeBank (+2.5 and +2.1 points on F1 value respectively).

show abstract

Section: B1 Dual Learningmentioning

confidence: 99%

LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification

Zuo¹,

Cao²,

Chen³

et al. 2021

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Confer

View full text Add to dashboard Cite

show abstract

“…The abundance of high-quality labeled data is critical for effectively training supervised models [ 31 , 51 ]. However, manually annotating NL utterances with their corresponding SC is expensive, cumbersome, and time consuming [ 10 , 38 , 52 , 53 ]. While emerging datasets have sprung up, due to the limitations in quantity, quality, and domain-crossing, the limited availability of labeled data is still becoming the primary bottleneck for data-driven supervised models [ 49 ].…”

Section: Insights Gained From Nlscg Research Backlogmentioning

confidence: 99%

“…Genres based on search concepts, including combining retrieval methods into neural code generation models [ 18 , 46 ], reranking a list of potential candidates [ 6 ], etc., are relatively feasible. Besides, data augmentation [ 45 , 63 ] processing based on training datasets , as well as the two-stage [ 52 ] and sketch [ 24 ]-based generation tactics, are also worthy of consideration.…”

Section: Perspectives On Nlscg Latest Technology Landscapementioning

confidence: 99%

“…The well formedness, type-sensitiveness, executability [ 4 ], and semantics of source code create the most difficulties. Many studies [ 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , ...…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Recent Advances in Intelligent Source Code Generation: A Survey on Natural Language Based Studies

Yang

Liu

Yin

2021

Entropy

View full text Add to dashboard Cite

Source Code Generation (SCG) is a prevalent research field in the automation software engineering sector that maps specific descriptions to various sorts of executable code. Along with the numerous intensive studies, diverse SCG types that integrate different scenarios and contexts continue to emerge. As the ultimate purpose of SCG, Natural Language-based Source Code Generation (NLSCG) is growing into an attractive and challenging field, as the expressibility and extremely high abstraction of the input end. The booming large-scale dataset generated by open-source code repositories and Q&A resources, the innovation of machine learning algorithms, and the development of computing capacity make the NLSCG field promising and give more opportunities to the model implementation and perfection. Besides, we observed an increasing interest stream of NLSCG relevant studies recently, presenting quite various technical schools. However, many studies are bound to specific datasets with customization issues, producing occasional successful solutions with tentative technical methods. There is no systematic study to explore and promote the further development of this field. We carried out a systematic literature survey and tool research to find potential improvement directions. First, we position the role of NLSCG among various SCG genres, and specify the generation context empirically via software development domain knowledge and programming experiences; second, we explore the selected studies collected by a thoughtfully designed snowballing process, clarify the NLSCG field and understand the NLSCG problem, which lays a foundation for our subsequent investigation. Third, we model the research problems from technical focus and adaptive challenges, and elaborate insights gained from the NLSCG research backlog. Finally, we summarize the latest technology landscape over the transformation model and depict the critical tactics used in the essential components and their correlations. This research addresses the challenges of bridging the gap between natural language processing and source code analytics, outlines different dimensions of NLSCG research concerns and technical utilities, and shows a bounded technical context of NLSCG to facilitate more future studies in this promising area.

show abstract

“…The dual learning mechanism enables a pair of dual systems to automatically learn from unlabeled data through a closed-loop game. The idea of dual learning has been applied into various tasks, such as Question Answer (Tang et al, 2017)/Generation (Tang et al, 2018), Image-to-Image Translation (Yi et al, 2017), Open-domain Information Extraction/Narration , Text Simplification , Semantic Parsing (Cao et al, 2019;Cao et al, 2020) and dialogue state tracking (Chen et al, 2020c).…”

Section: End-to-endmentioning

confidence: 99%

Decoupled Dialogue Modeling and Semantic Parsing for Multi-Turn Text-to-SQL

Chen¹,

Chen²,

Li³

et al. 2021

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021

Self Cite

View full text Add to dashboard Cite

Recently, Text-to-SQL for multi-turn dialogue has attracted great interest. Here, the user input of the current turn is parsed into the corresponding SQL query of the appropriate database, given all previous dialogue history. Current approaches mostly employ end-to-end models and consequently face two challenges. First, dialogue history modeling and Text-to-SQL parsing are implicitly combined, hence it is hard to carry out interpretable analysis and obtain targeted improvement. Second, SQL annotation of multi-turn dialogue is very expensive, leading to training data sparsity. In this paper, we propose a novel decoupled multi-turn Text-to-SQL framework, where an utterance rewrite model first explicitly solves completion of dialogue context, and then a single-turn Text-to-SQL parser follows. A dual learning approach is also proposed for the utterance rewrite model to address the data sparsity problem. Compared with end-to-end approaches, the proposed decoupled method can achieve excellent performance without any annotated in-domain data. With just a few annotated rewrite cases, the decoupled method outperforms the released state-of-the-art endto-end models on both SParC and CoSQL datasets.

show abstract

Unsupervised Dual Paraphrasing for Two-stage Semantic Parsing

Cited by 26 publications

References 34 publications

LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification

LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification

Recent Advances in Intelligent Source Code Generation: A Survey on Natural Language Based Studies

Decoupled Dialogue Modeling and Semantic Parsing for Multi-Turn Text-to-SQL

Contact Info

Product

Resources

About