Airbnb Introduces CASTLE for Training Natural Language Search Models with Synthetic Data
AI-generated summary Airbnb introduced CASTLE, a framework that uses large language models and de-identified booking sessions to generate synthetic search queries and relevance labels for natural language search models. In evaluations, CASTLE achieved a Kullback-Leibler divergence of 1.01 against real-user query distributions and received agreement ratings ranging from 91% to 93% from human evaluators.