Unsupervised Opinion Summarization as Copycat-Review Generation

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5151–5169 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 5151 Unsupervised Opinion Summarization as Copycat-Review Generation Arthur Braˇzinskas1, Mirella Lapata1 and Ivan Titov1,2 1 ILCC, University of Edinburgh 2 ILLC, University of Amsterdam abrazinskas@ed.ac.uk, {mlap,ititov}@inf.ed.ac.uk Abstract Opinion summarization is the task of automati- cally creating summaries that reflect subjective information expressed in multiple documents, such as product reviews. While the majority of previous work has focused on the extrac- tive setting, i.e., selecting fragments from in- put reviews to produce a summary, we let the model generate novel sentences and hence pro- duce abstractive summaries. Recent progress in summarization has seen the development of supervised models which rely on large quanti- ties of document-summary pairs. Since such training data is expensive to acquire, we in- stead consider the unsupervised setting, in other words, we do not use any summaries in training. We define a generative model for a review collection which capitalizes on the intu- ition that when generating a new review given a set of other reviews of a product, we should be able to control the “amount of novelty” go- ing into the new review or, equivalently, vary the extent to which it deviates from the input. At test time, when generating summaries, we force the novelty to be minimal, and produce a text reflecting consensus opinions. We capture this intuition by defining a hierarchical varia- tional autoencoder model. Both individual re- views and the products they correspond to are associated with stochastic latent codes, and the review generator (“decoder”) has direct access to the text of input reviews through the pointer- generator mechanism. Experiments on Ama- zon and Yelp datasets, show that setting at test time the review’s latent code to its mean, al- lows the model to produce fluent and coherent summaries reflecting common opinions. 1 Introduction Summarization of user opinions expressed in on- line resources, such as blogs, reviews, social media, or internet forums, has drawn much attention due to its potential for various information access appli- cations, such as creating digests, search, and report Summary This restaurant is a hidden gem in Toronto. The food is delicious, and the service is im- peccable. Highly recommend for anyone who likes French bistro. Reviews We got the steak frites and the chicken frites both of which were very good ... Great service ... || I really love this place ... Cˆote de Boeuf ... A Jewel in the big city ... || French jewel of Spadina and Adelaide , Jules ... They are super accommodating ... moules and frites are delicious ... || Food came with tons of greens and fries along with my main course , thumbs uppp ... || Chef has a very cool and fun attitude ... || Great little French Bistro spot ... Go if you want French bistro food classics ... || Great place ... the steak frites and it was amaz- ing ... Best Steak …

Content truncated to 500 words