Multimodal Customized Review Generation
Abstract: In this study, we introduce a new task called \emph{customized review generation}. This task aims to generate a personalized review that a specific user would give to a product that they have not yet reviewed. This can help users write high-quality reviews for products they have not previously reviewed, providing them with valuable insights. Additionally, customized reviews can offer a tailored summary of all reviews for a product, catering to the individual preferences of the reader.
To achieve this goal, we explore the use of multimodal information for customized review generation.
Specifically, we utilize a \emph{multimodal pre-trained language model} that takes a picture of a product and a set of words as input and generates a customized review using both visual and textual information.
Our experimental results demonstrate the effectiveness of the proposed model in generating customized reviews that are often of high quality.
Paper Type: Long
Research Area: Sentiment Analysis, Stylistic Analysis, and Argument Mining
Research Area Keywords: Sentiment Analysis
Languages Studied: English
Submission Number: 4316
Loading