Abstract
We introduce PRIMERA, a pre-trained model for multi-document representation with a focus on summarization that reduces the need for dataset-specific architectures and large amounts of fine-tuning labeled data. PRIMERA uses our newly proposed pre-training objective designed to teach the model to connect and aggregate information across documents. It also uses efficient encoder-decoder transformers to simplify the processing of concatenated input documents. With extensive experiments on 6 multi-document summarization datasets from 3 different domains on zero-shot, few-shot and full-supervised settings, PRIMERA outperforms current state-of-the-art dataset-specific and pre-trained models on most of these settings with large margins.- Anthology ID:
- 2022.acl-long.360
- Volume:
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- May
- Year:
- 2022
- Address:
- Dublin, Ireland
- Editors:
- Smaranda Muresan, Preslav Nakov, Aline Villavicencio
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 5245–5263
- Language:
- URL:
- https://aclanthology.org/2022.acl-long.360
- DOI:
- 10.18653/v1/2022.acl-long.360
- Cite (ACL):
- Wen Xiao, Iz Beltagy, Giuseppe Carenini, and Arman Cohan. 2022. PRIMERA: Pyramid-based Masked Sentence Pre-training for Multi-document Summarization. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5245–5263, Dublin, Ireland. Association for Computational Linguistics.
- Cite (Informal):
- PRIMERA: Pyramid-based Masked Sentence Pre-training for Multi-document Summarization (Xiao et al., ACL 2022)
- PDF:
- https://preview.aclanthology.org/emnlp-22-attachments/2022.acl-long.360.pdf
- Code
- allenai/primer + additional community code
- Data
- Arxiv HEP-TH citation graph, Multi-News, NewSHead, WCEP, WikiSum, arXiv Summarization Dataset