Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Identifying Shopping Intent in Product QA for Proactive Recommendations

Besnik Fetahu, Nachshon Cohen, Elad Haramaty, Liane Lewin-Eytan, Oleg Rokhlenko, Shervin Malmasi

TL;DR

This work tackles the challenge of identifying shopping need behind product questions in voice-enabled commerce to enable proactive recommendations. It introduces SPQI, a framework that fuses textual, product- and behavior-based signals via Mixture-of-Experts and uses Graph Attention Networks to propagate information across similarly themed questions. Using a large dataset of $374000$ PQs and a $28$-day purchase window for labeling, the approach achieves a high offline F1 of $0.91$ and demonstrates online benefits, including higher likelihood of adding items to shopping lists when SPQ prompts are shown. The results reveal the latent nature of shopping intent in voice interactions and show that relying on text alone is insufficient, highlighting the value of behavioral signals and graph-based contextual reasoning for proactive voice shopping experiences.

Abstract

Voice assistants have become ubiquitous in smart devices allowing users to instantly access information via voice questions. While extensive research has been conducted in question answering for voice search, little attention has been paid on how to enable proactive recommendations from a voice assistant to its users. This is a highly challenging problem that often leads to user friction, mainly due to recommendations provided to the users at the wrong time. We focus on the domain of e-commerce, namely in identifying Shopping Product Questions (SPQs), where the user asking a product-related question may have an underlying shopping need. Identifying a user's shopping need allows voice assistants to enhance shopping experience by determining when to provide recommendations, such as product or deal recommendations, or proactive shopping actions recommendation. Identifying SPQs is a challenging problem and cannot be done from question text alone, and thus requires to infer latent user behavior patterns inferred from user's past shopping history. We propose features that capture the user's latent shopping behavior from their purchase history, and combine them using a novel Mixture-of-Experts (MoE) model. Our evaluation shows that the proposed approach is able to identify SPQs with a high score of F1=0.91. Furthermore, based on an online evaluation with real voice assistant users, we identify SPQs in real-time and recommend shopping actions to users to add the queried product into their shopping list. We demonstrate that we are able to accurately identify SPQs, as indicated by the significantly higher rate of added products to users' shopping lists when being prompted after SPQs vs random PQs.

Identifying Shopping Intent in Product QA for Proactive Recommendations

TL;DR

This work tackles the challenge of identifying shopping need behind product questions in voice-enabled commerce to enable proactive recommendations. It introduces SPQI, a framework that fuses textual, product- and behavior-based signals via Mixture-of-Experts and uses Graph Attention Networks to propagate information across similarly themed questions. Using a large dataset of PQs and a -day purchase window for labeling, the approach achieves a high offline F1 of and demonstrates online benefits, including higher likelihood of adding items to shopping lists when SPQ prompts are shown. The results reveal the latent nature of shopping intent in voice interactions and show that relying on text alone is insufficient, highlighting the value of behavioral signals and graph-based contextual reasoning for proactive voice shopping experiences.

Abstract

Voice assistants have become ubiquitous in smart devices allowing users to instantly access information via voice questions. While extensive research has been conducted in question answering for voice search, little attention has been paid on how to enable proactive recommendations from a voice assistant to its users. This is a highly challenging problem that often leads to user friction, mainly due to recommendations provided to the users at the wrong time. We focus on the domain of e-commerce, namely in identifying Shopping Product Questions (SPQs), where the user asking a product-related question may have an underlying shopping need. Identifying a user's shopping need allows voice assistants to enhance shopping experience by determining when to provide recommendations, such as product or deal recommendations, or proactive shopping actions recommendation. Identifying SPQs is a challenging problem and cannot be done from question text alone, and thus requires to infer latent user behavior patterns inferred from user's past shopping history. We propose features that capture the user's latent shopping behavior from their purchase history, and combine them using a novel Mixture-of-Experts (MoE) model. Our evaluation shows that the proposed approach is able to identify SPQs with a high score of F1=0.91. Furthermore, based on an online evaluation with real voice assistant users, we identify SPQs in real-time and recommend shopping actions to users to add the queried product into their shopping list. We demonstrate that we are able to accurately identify SPQs, as indicated by the significantly higher rate of added products to users' shopping lists when being prompted after SPQs vs random PQs.
Paper Structure (17 sections, 1 equation, 2 figures, 1 table)

This paper contains 17 sections, 1 equation, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Data Analysis
  4. Dataset and Labels
  5. Data Insights
  6. Questions.
  7. Users.
  8. Identifying SPQs
  9. Learning Joint Question Representations via MoE
  10. Question Graph Construction & Classification
  11. Experiments
  12. Baselines and Ablations
  13. Evaluation Results
  14. Offline Evaluation
  15. Model Performance.
  16. ...and 2 more sections

Figures (2)

  • Figure 1: User A searches for ingredients for a recipe. Our approach identifies the user having a shopping need, hence, triggering voice assistant's shopping recommendations. For User B's (likely vegetarian) our model predicts no shopping intent, hence, after answering, the turn is over.
  • Figure 2: SPQ rate per product type (out of 50 types with at least 1000 PQs, which represent 29.7% of the entire dataset).