Discrete Optimization for Unsupervised Sentence Summarization with Word-Level Extraction

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5032–5042 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 5032 Discrete Optimization for Unsupervised Sentence Summarization with Word-Level Extraction Raphael Schumann1, Lili Mou2, Yao Lu3, Olga Vechtomova3, Katja Markert1 1Institute of Computational Linguistics, Heidelberg University, Germany {rschuman, markert}@cl.uni-heidelberg.de 2University of Alberta, Canada; Alberta Machine Intelligence Institute (Amii) doublepower.mou@gmail.com 3University of Waterloo, Canada {yao.lu, ovechtom}@uwaterloo.ca Abstract Automatic sentence summarization produces a shorter version of a sentence, while preserv- ing its most important information. A good summary is characterized by language fluency and high information overlap with the source sentence. We model these two aspects in an unsupervised objective function, consisting of language modeling and semantic similarity metrics. We search for a high-scoring sum- mary by discrete optimization. Our proposed method achieves a new state-of-the art for un- supervised sentence summarization according to ROUGE scores. Additionally, we demon- strate that the commonly reported ROUGE F1 metric is sensitive to summary length. Since this is unwillingly exploited in recent work, we emphasize that future evaluation should explic- itly group summarization systems by output length brackets.1 1 Introduction Sentence summarization transforms a long source sentence into a short summary, while preserving key information (Rush et al., 2015). Sentence summarization has wide applications, for example, news headline generation and text simplification. State-of-the-art sentence summarization systems are based on sequence-to-sequence neural net- works (Rush et al., 2015; Nallapati et al., 2016; Wang et al., 2019), which require massive parallel data for training. Therefore, unsupervised sentence summarization has recently attracted increasing interest. Cycle-consistency approaches treat the summary as a discrete latent variable and use it to reconstruct the source sentence (Wang and Lee, 2018; Baziotis et al., 2019). Such latent-space gen- eration fails to explicitly model the resemblance be- tween the source sentence and the target summary. 1Our code and system outputs are available at: https://github.com/raphael-sch/HC_ Sentence_Summarization the world 's biggest miner bhp billiton announced tuesday it was dropping its controversial hostile takeover bid for rival rio tinto due to the state of the global economy bhp billiton dropping hostile bid for rio tinto bhp billiton drops rio tinto takeover bid summary: reference: Figure 1: Summarizing a sentence x by hill climbing. Each row is a Boolean vector at at a search step t . A black cell indicates a word is selected, and vice versa. Randomly swapping two values in the Boolean vector yields a new summary that is scored by an objective function that measures language fluency and semantic similarity. If the new summary increases the objective, this summary is accepted as the current best solution. Rejected solutions are not depicted. Zhou and Rush (2019) propose a left-to-right beam search approach based on a heuristically defined scoring function. However, beam search is biased towards the first few words of the source. In this paper, we propose a hill-climbing ap- proach to unsupervised sentence summarization, directly extracting words from the source sentence. This is motivated by the observation that human- written …

Content truncated to 500 words