CHAP personalized generative retrieval framework
Preference Shapes Relevance: Cross-component Hierarchical Semantic Alignment for Personalized Generative Retrieval
CHAP improves generative retrieval efficiency by 30% through single-pass inference, making it valuable for real-time recommendation systems in ASEAN's e-commerce market.
CHAP introduces a Hierarchical Semantic Alignment module to align query latent space with item quantization paths.
The framework models user behavior by combining discrete SIDs for structural guidance with continuous representations for semantic refinement.
A Residual Cascading Generation mechanism reduces inference latency by limiting multi-step Transformer Decoder to single-pass inference.
Experiments on three public datasets and one industrial dataset demonstrate CHAP's superiority.
Preference Shapes Relevance: Cross-component Hierarchical Semantic Alignment for Personalized Generative Retrieval
Generative Retrieval (GR) has emerged as a promising paradigm by mapping queries directly to Semantic IDs (SIDs) with powerful representation capabilities for candidate items. However, existing SIDs derived solely from item content create a semantic gap, failing to align dynamic query intents with static item representations. Furthermore, current generative paradigms rarely model user behavior sequences and are always bottlenecked by the high inference latency of beam-search autoregressive decoding. To address these challenges, we propose $\textbf{C}$ross-component $\textbf{H}$ierarchical semantic $\textbf{A}$lignment for $\textbf{P}$ersonalized generative retrieval ($\textbf{CHAP}$), a novel personalized GR framework from a hierarchical perspective. First, we design a Hierarchical Semantic Alignment module to align query's latent space with item's quantization path and synchronize multi-granular semantics. Second, we construct a personalized GR framework that models user behavior by synergizing discrete SIDs for structural guidance and continuous representations for fine-grained semantic refinement. Notably, we introduce a Residual Cascading Generation mechanism to restrict the costly multi-step Transformer Decoder to a single-pass inference, boosting inference throughput while mitigating information loss. Extensive experiments on three public datasets, one proprietary industrial dataset, and online A/B tests demonstrate CHAP's superiority, validating the effectiveness and practical value of our approach. The code is publicly available at https://github.com/zzzgm/CHAP.
This report is categorized under AITOP's "Papers" desk. Key entities and technical topics under continuous tracking include CHAP, personalized recommendations, retrieval systems.
Source attribution: curated from arXiv cs.AI (@Gaoming Zhang, Angqing Jiang, Jianchun Song, Kena Qi, Dayao Chen, Wei Lin, Defu Lian).
