A Tree-to-Sequence Model for Neural NLG in Task-Oriented Dialog

By Jinfeng Rao, Kartikeya Upasani, Anusha Balakrishnan, Michael. White, Anuj Kumar, Rajen Subba. 2019

14 citations recorded by the source corpus.

Generating fluent natural language responses from structured semantic representations is a critical step in task-oriented conversational systems. Sequence-to-sequence models on flat meaning representations (MR) have been dominant in this task, for example in the E2E NLG Challenge. Previous work has shown that a tree-structured MR can improve the model for better discourse-level structuring and sentence-level planning. In this work, we propose a tree-to-sequence model that uses a tree-LSTM encoder to leverage the tree structures in the input MR, and further enhance the decoding by a structure-enhanced attention mechanism. In addition, we explore combining these enhancements with constrained decoding to improve semantic correctness. Our experiments not only show significant improvements over standard seq2seq baselines, but also is more data-efficient and generalizes better to hard scenarios.

Subfields: Artificial Intelligence, Natural Language Processing

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 Processing

Citations in the corpus

Structure-to-Text Generation with Self-Training, Acceptability Classifiers and Context-Conditioning for the GEM Shared Task (2021, background): “In the research community, generative models (Wen et al., 2015; Duˇsek and Jurcıcek, 2016; Rao et al., 2019) have become popular for their data-driven scaling story and superior naturalness over typical template-based systems (Gatt and Krahmer, 2018; Dale, 2020).”
Improving Cross-Modal Alignment in Vision Language Navigation via Syntactic Information (2021, background): “, 2016), dialogue (Rao et al., 2019), and language inference (Chen et al.”
AggGen: Ordering and Aggregating while Generating (2021, methodology): “For applications where structured inputs are present, neural graph encoders (Marcheggiani and Perez-Beltrachini, 2018; Rao et al., 2019; Gao et al., 2020) or decoding of explicit graph references (Logan et al., 2019) are applied for higher accuracy.”
Getting to Production with Few-shot Natural Language Generation Models (2021, background): “Recently, generative models (Wen et al., 2015; Dušek and Jurcıcek, 2016; Rao et al., 2019) have become popular for their data-driven scaling story and superior naturalness over the typical templatebased systems (Gatt and Krahmer, 2018; Dale,”