From IDs to Semantics: A Generative Framework for Cross-Domain Recommendation with Adaptive Semantic Tokenization

2025/11/11 by Peiyu Hu, Hu, Peiyu, Wen Feng Lu +3 · 3 citations
Computer Science · #Advanced Graph Neural Networks #FOS: Computer and information sciences #Information Retrieval (cs.IR) #Machine Learning in Healthcare #Recommender Systems and Techniques

paper · pdf · doi:10.48550/arxiv.2511.08006

openalex publication_date 2025/11/11 · openalex created_date 2025/11/13 · openalex updated_date 2026/07/28

Abstract

Cross-domain recommendation (CDR) is crucial for improving recommendation accuracy and generalization, yet traditional methods are often hindered by the reliance on shared user/item IDs, which are unavailable in most real-world scenarios. Consequently, many efforts have focused on learning disentangled representations through multi-domain joint training to bridge the domain gaps. Recent Large Language Model (LLM)-based approaches show promise, they still face critical challenges, including: (1) the item ID tokenization dilemma, which leads to vocabulary explosion and fails to capture high-order collaborative knowledge; and (2) insufficient domain-specific modeling for the complex evolution of user interests and item semantics. To address these limitations, we propose GenCDR, a novel Generative Cross-Domain Recommendation framework. GenCDR first employs a Domain-adaptive Tokenization module, which generates disentangled semantic IDs for items by dynamically routing between a universal encoder and domain-specific adapters. Symmetrically, a Cross-domain Autoregressive Recommendation module models user preferences by fusing universal and domain-specific interests. Finally, a Domain-aware Prefix-tree enables efficient and accurate generation. Extensive experiments on multiple real-world datasets demonstrate that GenCDR significantly outperforms state-of-the-art baselines. Our code is available in the supplementary materials.

Citations

• Multi-Modal Multi-Behavior Sequential Recommendation with Conditional Diffusion-Based Feature Denoising
• HiGarment: Cross-modal Harmony Based Diffusion Model for Flat Sketch to Realistic Garment Image
• Towards Cross-modal Retrieval in Chinese Cultural Heritage Documents: Dataset and Solution
• Bridge the Domains: Large Language Models Enhanced Cross-domain Sequential Recommendation
• ActionPiece: Contextually Tokenizing Action Sequences for Generative Recommendation
• UniHR: Hierarchical Representation Learning for Unified Knowledge Graph Link Prediction
• Mitigating Propensity Bias of Large Language Models for Recommender Systems
• LLMEmb: Large Language Model Can Be a Good Embedding Generator for Sequential Recommendation
• Large Language Models Enhanced Collaborative Filtering
• Post-Training Attribute Unlearning in Recommender Systems
• Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations
• Adapting Large Language Models by Integrating Collaborative Semantics for Recommendation
• Bridging Items and Language: A Transition Paradigm for Large Language Model-Based Recommendation
• A Survey on Large Language Models for Recommendation
• Learning Vector-Quantized Item Representation for Transferable Sequential Recommenders
• Autoregressive Image Generation using Residual Quantization
• Sequential Recommendation via Stochastic Self-Attention
• BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer
• Self-Attentive Sequential Recommendation
• Neural Discrete Representation Learning
• Image-based Recommendations on Styles and Substitutes
• Improving LLM-powered Recommendations with Personalized Information

Cited by

• Token Factory: Efficiently Integrating Diverse Signals into Large Recommendation Models
• A Hierarchical Quantized Tokenization Framework for Task-Adaptive Graph Representation Learning
• Hierarchical Latent Reasoning for LLM-based Recommendation
• Empowering Cross-Domain Sequential Recommendation with Hybrid Tokenization and Serial-Parallel Decoding

Related

• Generative Multi-Target Cross-Domain Recommendation
• ID-centric Pre-training for Recommendation
• Tokenize Once, Recommend Anywhere: Unified Item Tokenization for Multi-domain LLM-based Recommendation
• Break the ID-Language Barrier: An Adaption Framework for LLM-based Sequential Recommendation
• Generative Recommendation with Semantic IDs: A Practitioner's Handbook