Problem
Research questions and friction points this paper is trying to address.
query intent
e-commerce search
under-specified queries
user intent detection
ambiguous queries
IntentTune: Using user demand and personalization to resolve "unknown" query intents for e-commerce search
🤖 AI Summary
This work addresses the challenge of intent recognition in e-commerce search for ambiguous queries—such as those lacking explicit attributes like gender or age—by proposing an end-to-end parsing framework that integrates personalized user behavioral signals (e.g., search and browsing history) with population-level demand patterns. The approach jointly models individual user behavior and aggregated group preferences, enabling multi-granular intent classification. It provides the first systematic validation of the significant advantages conferred by user behavior signals in inferring ambiguous intents. Experimental results demonstrate that the proposed framework substantially outperforms methods relying solely on population-level statistics or static user profiles across key intent dimensions, including gender, age, product category, and size.
📝 Abstract
Understanding user intent is fundamental to delivering relevant search results in e-commerce. However, substantial fraction of real-world queries are under-specified (e.g., "watch" or "shirt"), lacking explicit attributes such as gender or age group. This ambiguity poses a significant challenge for query intent detection models in e-commerce search systems, which must accurately infer latent user intent (e.g., age, gender) to support effective downstream retrieval. We introduce IntentTune, a framework for resolving ambiguous or under-specified query intents by leveraging either (1) user-specific behavioral signals including search history, browsing activity, and profile attributes or (2) population-level demand patterns aggregated across all users. Through experiments on real-world e-commerce data, we first demonstrate that population-level demand patterns alone are insufficient to reliably infer intent in under-specified queries. We then demonstrate that user-specific behavioral signals -- particularly prior search queries -- outperform both population-level statistics and static profile information for inferring gender, age group, product category, and size intent from underspecified queries.
Innovation
Methods, ideas, or system contributions that make the work stand out.
intent resolution
personalization
user behavior signals
e-commerce search
under-specified queries
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
