Question-Answer Selection in User to User Marketplace Conversations
This work addresses a domain-specific issue for sellers in online marketplaces, presenting an incremental improvement by applying existing neural methods to a new dataset.
The paper tackles the problem of sellers being overwhelmed by buyer questions in user-to-user marketplaces by developing a neural-network ranking model that selects relevant sentences from product descriptions to answer questions, using a dataset of 590K questions and answers.
Sellers in user to user marketplaces can be inundated with questions from potential buyers. Answers are often already available in the product description. We collected a dataset of around 590K such questions and answers from conversations in an online marketplace. We propose a question answering system that selects a sentence from the product description using a neural-network ranking model. We explore multiple encoding strategies, with recurrent neural networks and feed-forward attention layers yielding good results. This paper presents a demo to interactively pose buyer questions and visualize the ranking scores of product description sentences from live online listings.