Do We Really Need That Many Parameters In Transformer For Extractive Summarization? Discourse Can Help !

TL;DR

This paper introduces a discourse-informed, parameter-efficient self-attention mechanism for extractive summarization, achieving comparable or better performance than traditional transformer models with fewer parameters.

Contribution

It proposes a novel tree self-attention mechanism based on discourse priors, extending the Synthesizer framework for more parameter-efficient NLP models.

Findings

Achieves competitive ROUGE scores with fewer parameters.

Outperforms 8-head transformer on sentence-level summarization.

Reduces parameters significantly while maintaining performance.

Abstract

The multi-head self-attention of popular transformer models is widely used within Natural Language Processing (NLP), including for the task of extractive summarization. With the goal of analyzing and pruning the parameter-heavy self-attention mechanism, there are multiple approaches proposing more parameter-light self-attention alternatives. In this paper, we present a novel parameter-lean self-attention mechanism using discourse priors. Our new tree self-attention is based on document-level discourse information, extending the recently proposed "Synthesizer" framework with another lightweight alternative. We show empirical results that our tree self-attention approach achieves competitive ROUGE-scores on the task of extractive summarization. When compared to the original single-head transformer model, the tree attention approach reaches similar performance on both, EDU and sentence level, despite the significant reduction of parameters in the attention component. We further significantly outperform the 8-head transformer model on sentence level when applying a more balanced hyper-parameter setting, requiring an order of magnitude less parameters.