Improving Product Relevance for Online Voice Groceries

Walmart Labs Voice Computer Science, 2019-20

Liaison(s): Ghodrat Aalipour, Venkata Nandanavanam
Advisor(s): Colleen Lewis (F), Ran Libeskind-Hadas (S)
Students(s): Shannon Steele (PM-S), Ka Ki Fung (PM-F), Richy Chen, Justin Gadalla

The purpose of Walmart Labs’ Clinic project is to improve the product relevance for online voice grocery queries. There are two components of the project: feature extraction for concepts and question generation for Walmart’s question answering system. For feature extraction, we algorithmically extract candidate features from user’s search queries. To expand the Q-and-A system, we generate generic question templates from user-submitted questions, then for specific concepts, we can use the generic templates and populate them with extracted product information and candidate features.