2019/10/18 by Manirupa Das, Zhen Wang, Das, Manirupa +9
Computer Science · #Computation and Language (cs.CL) #Expert finding and Q&A systems #FOS: Computer and information sciences #Information Retrieval (cs.IR) #Machine Learning (cs.LG) #Sentiment Analysis and Opinion Mining #Topic Modeling
paper · pdf · doi:10.48550/arxiv.1910.08270
openalex publication_date 2019/10/18 · openalex created_date 2022/07/28 · openalex updated_date 2026/07/28
Online customer reviews on large-scale e-commerce websites, represent a rich\nand varied source of opinion data, often providing subjective qualitative\nassessments of product usage that can help potential customers to discover\nfeatures that meet their personal needs and preferences. Thus they have the\npotential to automatically answer specific queries about products, and to\naddress the problems of answer starvation and answer augmentation on associated\nconsumer Q & A forums, by providing good answer alternatives. In this work, we\nexplore several recently successful neural approaches to modeling sentence\npairs, that could better learn the relationship between questions and ground\ntruth answers, and thus help infer reviews that can best answer a question or\naugment a given answer. In particular, we hypothesize that our adversarial\ndomain adaptation-based approach, due to its ability to additionally learn\ndomain-invariant features from a large number of unlabeled, unpaired\nquestion-review samples, would perform better than our proposed baselines, at\nanswering specific, subjective product-related queries using reviews. We\nvalidate this hypothesis using a small gold standard dataset of question-review\npairs evaluated by human experts, significantly surpassing our chosen\nbaselines. Moreover, our approach, using no labeled question-review sentence\npair data for training, gives performance at par with another method utilizing\nlabeled question-review samples for the same task.\n