KuaiSearch: A Large-Scale E-Commerce Search Dataset for Recall, Ranking, and Relevance
This massive real-world dataset could finally fix terrible online shopping search results.
Deep Dive
Researchers from Kuaishou have released KuaiSearch, reportedly the largest e-commerce search dataset available. Built from real user interactions on the Kuaishou platform, it contains authentic queries and product texts, includes cold-start users and long-tail products, and spans the three key stages of the search pipeline: recall, ranking, and relevance. The dataset aims to address limitations in existing resources and advance LLM-based search research with more realistic benchmarks.
Why It Matters
Better-trained AI models using this data could dramatically improve product discovery and search accuracy for billions of online shoppers.