Generating EDU Extracts for Plan-Guided Summary Re-Ranking
Preprint
Abstract
Two-step approaches, in which summary candidates are generated-then-reranked to return a single summary, can improve ROUGE scores over the standard single-step approach. Yet, standard decoding methods (i.e., beam search, nucleus sampling, and diverse beam search) produce candidates with redundant, and often low quality, content. In this paper, we design a novel method to generate candidates for re-ranking that addresses these issues. We ground each candidate abstract on its own unique content plan and generate distinct plan-guided abstracts using a model's top beam. More concretely, a standard language model (a BART LM) auto-regressively generates elemental discourse unit (EDU) content plans with an extractive copy mechanism. The top K beams from the content plan generator are then used to guide a separate LM, which produces a single abstractive candidate for each distinct plan. We apply an existing re-ranker (BRIO) to abstractive candidates generated from our method, as well as baseline decoding methods. We show large relevance improvements over previously published methods on widely used single document news article corpora, with ROUGE-2 F1 gains of 0.88, 2.01, and 0.38 on CNN / Dailymail, NYT, and Xsum, respectively. A human evaluation on CNN / DM validates these results. Similarly, on 1k samples from CNN / DM, we show that prompting GPT-3 to follow EDU plans outperforms sampling-based methods by 1.05 ROUGE-2 F1 points. Code to generate and realize plans is available at https://github.com/griff4692/edu-sum.Keywords
All Related Versions
This publication has 41 references indexed in Scilit:
- Benchmarking Large Language Models for News SummarizationTransactions of the Association for Computational Linguistics, 2024
- Repairing the Cracked Foundation: A Survey of Obstacles in Evaluation Practices for Generated TextJournal of Artificial Intelligence Research, 2023
- Planning with Learned Entity Prompts for Abstractive SummarizationTransactions of the Association for Computational Linguistics, 2021
- On Extractive and Abstractive Neural Document Summarization with Transformer Language Models2020
- Learning to Fuse Sentences with Transformers for Summarization2020
- BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension2020
- Unsupervised Cross-lingual Representation Learning at Scale2020
- Exploring Content Selection in Summarization of Novel Chapters2020
- Composing Elementary Discourse Units in Abstractive Summarization2020
- Diverse Beam Search for Improved Description of Complex ScenesProceedings of the AAAI Conference on Artificial Intelligence, 2018