# Generating Summaries with Topic Templates and Structured Convolutional Decoders

## TL;DR

This paper introduces a structured convolutional decoder guided by content structure to improve multi-sentence text summaries, outperforming existing models in content coverage across multiple domains.

## Contribution

It proposes a novel structured convolutional decoder that incorporates content structure guidance, enhancing summary quality over traditional sequential decoders.

## Key findings

- Better content coverage in summaries
- Improved performance on multiple datasets
- Validated by automatic and human evaluations

## Abstract

Existing neural generation approaches create multi-sentence text as a single sequence. In this paper we propose a structured convolutional decoder that is guided by the content structure of target summaries. We compare our model with existing sequential decoders on three data sets representing different domains. Automatic and human evaluation demonstrate that our summaries have better content coverage.