Text Summarization as Tree Transduction by Top-Down TreeLSTM

TL;DR

This paper presents a novel neural approach to extractive text summarization by modeling it as a parse tree transduction task using a specialized TreeLSTM, achieving state-of-the-art results.

Contribution

It introduces a new tree transduction framework for extractive compression, extending TreeLSTM to incorporate parent-child structural relationships.

Findings

Achieves state-of-the-art performance on sentence compression benchmarks.

Outperforms previous models in accuracy and compression rate.

Models structural relationships effectively in tree transduction.

Abstract

Extractive compression is a challenging natural language processing problem. This work contributes by formulating neural extractive compression as a parse tree transduction problem, rather than a sequence transduction task. Motivated by this, we introduce a deep neural model for learning structure-to-substructure tree transductions by extending the standard Long Short-Term Memory, considering the parent-child relationships in the structural recursion. The proposed model can achieve state of the art performance on sentence compression benchmarks, both in terms of accuracy and compression rate.