06-10, 16:30–17:10 (Europe/Berlin), Palais Atelier
Dense Vector Search is not the only route to improve your search relevance. Empower your existing sparse keyword search with semantic search capabilities by leveraging text expansion, metadata enrichment, and query re-writing techniques.
This presentation discusses about advanced methods to enhance keyword search by leveraging machine learning and large language models. The techniques involve utilising sparse models for document and query expansion, implementing metadata enrichment for text and images, and incorporating query rewriting using Large Language models. Through a live demonstration on a retail search scenario, we will illustrate how these approaches provide more relevant results by addressing the challenges of keyword search, such as vocabulary mismatches, unmatched queries against document attributes, and non-contextual search. The goal is to empower your existing keyword search system with semantic search capabilities without totally transforming to a costly and un-explainable, dense vector search system.
A search enthusiast actively researching and experimenting on using Machine Learning to improve relevance.
Hajer Bouafif is a solutions architect in Data Analytics and search with a background in Big Data engineering. Hajer provides organizations with best practices and well-architected reviews to build large-scale Machine Learning search solutions.