Tree-Structured Neural Machine for Linguistics-Aware Sentence Generation

TL;DR

This paper introduces a tree-structured neural model for linguistically-aware sentence generation, leveraging dependency trees to improve response relevance and fluency in conversational AI.

Contribution

It develops a novel tree-structured decoder and a tree canonicalization method to incorporate linguistic structure into sentence generation.

Findings

Outperforms baseline models with over 11% higher acceptance ratio

Uses dependency parsing to enhance linguistic relevance

Employs a tree-structured search for more probable responses

Abstract

Different from other sequential data, sentences in natural language are structured by linguistic grammars. Previous generative conversational models with chain-structured decoder ignore this structure in human language and might generate plausible responses with less satisfactory relevance and fluency. In this study, we aim to incorporate the results from linguistic analysis into the process of sentence generation for high-quality conversation generation. Specifically, we use a dependency parser to transform each response sentence into a dependency tree and construct a training corpus of sentence-tree pairs. A tree-structured decoder is developed to learn the mapping from a sentence to its tree, where different types of hidden states are used to depict the local dependencies from an internal tree node to its children. For training acceleration, we propose a tree canonicalization method, which transforms trees into equivalent ternary trees. Then, with a proposed tree-structured search method, the model is able to generate the most probable responses in the form of dependency trees, which are finally flattened into sequences as the system output. Experimental results demonstrate that the proposed X2Tree framework outperforms baseline methods over 11.15% increase of acceptance ratio.