Structured semantic supervision improves product search relevance
Beyond Relevance: Structured Semantic Supervision for Product Search with LLM-Augmented Annotations
Information RetrievalComputation and Language
Summary
Finding the exact product a shopper wants online is hard because many products are only somewhat related to what they type. This paper shows that adding detailed, structured information generated by large language models (LLMs) about queries and products helps search systems understand what people really want. The researchers combined these AI-generated signals with human checks and found their method made product search more accurate. They also found that human-edited explanations added the most value, meaning the AI helps show useful details rather than replace humans completely.
What this means in practice
- •For e-commerce engineers: Enhance product search by integrating structured semantic signals to better match user intent and improve ranking quality, especially for difficult queries.
- •For customer support platforms: Use structured semantic understanding to improve recommendation quality in chatbots or virtual assistants that help users find products precisely.
Authors
Girish A. Koushik, Swapnil Bhosale, Samarth Agrawal, Hadeel Sadany, Constantin Orasan, Xiatian Zhu, Diptesh Kanojia
Abstract
E-commerce search requires distinguishing products that are merely related to a query from those that directly satisfy the user's shopping intent. We augment query-product pairs with structured LLM-generated query and product attributes and human-validated relevance, explanations, and centrality judgments, and evaluate these signals using a simple dual-encoder retriever and MLP re-ranker. On an augmented subset of ESCI, a human-feature oracle reaches $0.9382$ nDCG@10, while a human-free trained $Q+P$ configuration reaches $0.9258$. Synthetic approximations of the human signals reach $0.9150$ overall but provide substantial gains for difficult, low-performing queries. Ablations show that most of the oracle improvement comes from post-edited explanations and annotator comments rather than the scalar centrality feature, suggesting that LLMs are most useful for exposing and approximating structured semantic supervision rather than replacing human judgment directly.