A Deep Reinforced Model for Zero-Shot Cross-Lingual Summarization with Bilingual Semantic Similarity Rewards

TL;DR

This paper introduces an end-to-end deep reinforcement learning model for zero-shot cross-lingual summarization, optimizing bilingual semantic similarity to improve summary quality without relying on machine translation pipelines.

Contribution

It presents a novel reinforcement learning approach that directly optimizes bilingual semantic similarity, enhancing cross-lingual summarization performance without extensive data or translation errors.

Findings

Reinforcement learning improves summary fluency and relevance.

The model performs well on English-Chinese and English-German tasks.

Bilingual semantic similarity rewards lead to better summaries.

Abstract

Cross-lingual text summarization aims at generating a document summary in one language given input in another language. It is a practically important but under-explored task, primarily due to the dearth of available data. Existing methods resort to machine translation to synthesize training data, but such pipeline approaches suffer from error propagation. In this work, we propose an end-to-end cross-lingual text summarization model. The model uses reinforcement learning to directly optimize a bilingual semantic similarity metric between the summaries generated in a target language and gold summaries in a source language. We also introduce techniques to pre-train the model leveraging monolingual summarization and machine translation objectives. Experimental results in both English--Chinese and English--German cross-lingual summarization settings demonstrate the effectiveness of our methods. In addition, we find that reinforcement learning models with bilingual semantic similarity as rewards generate more fluent sentences than strong baselines.