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.- Anthology ID:
- 2020.codi-1.13
- Volume:
- Proceedings of the First Workshop on Computational Approaches to Discourse
- Month:
- November
- Year:
- 2020
- Address:
- Online
- Editors:
- Chloé Braud, Christian Hardmeier, Junyi Jessy Li, Annie Louis, Michael Strube
- Venue:
- CODI
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 124–134
- Language:
- URL:
- https://aclanthology.org/2020.codi-1.13
- DOI:
- 10.18653/v1/2020.codi-1.13
- Cite (ACL):
- Wen Xiao, Patrick Huber, and Giuseppe Carenini. 2020. Do We Really Need That Many Parameters In Transformer For Extractive Summarization? Discourse Can Help !. In Proceedings of the First Workshop on Computational Approaches to Discourse, pages 124–134, Online. Association for Computational Linguistics.
- Cite (Informal):
- Do We Really Need That Many Parameters In Transformer For Extractive Summarization? Discourse Can Help ! (Xiao et al., CODI 2020)
- PDF:
- https://preview.aclanthology.org/ingest-2024-clasp/2020.codi-1.13.pdf
- Data
- CNN/Daily Mail