A Generative Framework for Personalized Sticker Retrieval

2025-09-22

AI Summary

Existing generative retrieval methods lack effective personalization modeling, limiting their ability to capture users’ diverse sticker preferences. To address this, we propose PEARL—the first generative framework for personalized sticker retrieval. PEARL (1) constructs a user representation model that jointly encodes click history and demographic/behavioral profile attributes; (2) introduces an intent-aware autoregressive generation objective to explicitly model high-level user intent; and (3) integrates multi-task learning with intent ranking to jointly optimize generation fidelity and relevance. Extensive offline evaluations and large-scale online A/B tests demonstrate that PEARL significantly outperforms state-of-the-art baselines: it improves retrieval accuracy by 12.7% and user click-through rate by 9.3%. These results validate PEARL’s dual advantages in both relevance and personalization, establishing a new benchmark for generative personalized retrieval.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Formulating information retrieval as a variant of generative modeling, specifically using autoregressive models to generate relevant identifiers for a given query, has recently attracted considerable attention. However, its application to personalized sticker retrieval remains largely unexplored and presents unique challenges: existing relevance-based generative retrieval methods typically lack personalization, leading to a mismatch between diverse user expectations and the retrieved results. To address this gap, we propose PEARL, a novel generative framework for personalized sticker retrieval, and make two key contributions: (i) To encode user-specific sticker preferences, we design a representation learning model to learn discriminative user representations. It is trained on three prediction tasks that leverage personal information and click history; and (ii) To generate stickers aligned with a user's query intent, we propose a novel intent-aware learning objective that prioritizes stickers associated with higher-ranked intents. Empirical results from both offline evaluations and online tests demonstrate that PEARL significantly outperforms state-of-the-art methods.

Problem

• Personalized sticker retrieval lacks existing generative methods
• Current approaches mismatch user expectations without personalization
• PEARL framework encodes preferences and aligns with intent

Innovation

• Generative framework for personalized sticker retrieval
• Representation learning model for user preferences
• Intent-aware learning objective for query alignment