End-to-End Conversational Search for Online Shopping with Utterance Transfer

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3477–3486 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 3477 End-to-End Conversational Search for Online Shopping with Utterance Transfer Liqiang Xiao1∗, Jun Ma2, Xin Luna Dong2, Pascual Martinez-Gomez2, Nasser Zalmout2, Chenwei Zhang2, Tong Zhao2, Hao He1,3, Yaohui Jin1,3 1MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University 2Amazon.com, Inc. 3State Key Lab of Advanced Optical Communication System and Network, Shanghai Jiao Tong University xiaoliqiang@sjtu.edu.cn, {junmaa, lunadong, gomepasc, nzalmout, cwzhang, zhaoton}@amazon.com, {hehao, jinyh}@sjtu.edu.cn Abstract Successful conversational search systems can present natural, adaptive and interactive shop- ping experience for online shopping cus- tomers. However, building such systems from scratch faces real word challenges from both imperfect product schema/knowledge and lack of training dialog data. In this work we first propose ConvSearch, an end-to-end con- versational search system that deeply com- bines the dialog system with search. leverages the text profile to retrieve products,It which is more robust against imperfect prod- uct schema/knowledge compared with using product attributes alone. We then address the lack of data challenges by proposing an utter- ance transfer approach that generates dialogue utterances by using existing dialog from other domains, and leveraging the search behavior data from e-commerce retailer. With utterance transfer, we introduce a new conversational search dataset for online shopping. Experi- ments show that our utterance transfer method can significantly improve the availability of training dialogue data without crowd-sourcing, and the conversational search system signifi- cantly outperformed the best tested baseline. 1 Introduction Search systems play significant roles in today’s online shopping experience. In conventional e- commerce search systems, user interacts with the system through typing of keywords, followed by product clicks or keywords modifications, depend- ing on whether returned product list matches with user expectation. The recent success of intelligent assistants such as Alexa, Google Now, and Siri enables user to interact with search systems using natural language. For online shopping in particular, ∗Work performed during internship at Amazon. it becomes alluring that users can navigate through products with conversations like traditional in-store shopping, guided by a knowledgeable yet thought- ful virtual shopping assistant. However, building a successful conversational search system for online shopping faces at least two real world challenges. The first challenge is the imperfect product attribute schema and product knowledge. While this challenge applies also to tra- ditional search systems, it is more problematic for conversational search because the later depends on product attributes to link lengthy multi-turn utter- ances (in contrast to short queries in conventional search) with products. Most previous conversa- tional shopping search work (Li et al., 2018a; Bi et al., 2019; Yan et al., 2017) looks for the target product through direct attribute matching, assum- ing availability of complete product knowledge in structured form. In practice, this assumption rarely holds, and systems designed with this assumption will suffer from product recall losses. The second challenge is the lack of in-domain dialog dataset for model training. Constructing a large-scale dialog dataset by crowd-sourcing …

Content truncated to 500 words