Saved in:
Bibliographic Details
Main Authors: Lin, Yuzhen, Chen, Hongyi, Chen, Xuanjing, Wang, Shaowen, Xu, Ivonne, Jiang, Dongming
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.21543
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908731816542208
author Lin, Yuzhen
Chen, Hongyi
Chen, Xuanjing
Wang, Shaowen
Xu, Ivonne
Jiang, Dongming
author_facet Lin, Yuzhen
Chen, Hongyi
Chen, Xuanjing
Wang, Shaowen
Xu, Ivonne
Jiang, Dongming
contents Generative recommendation models often struggle with two key challenges: (1) the superficial integration of collaborative signals, and (2) the decoupled fusion of multimodal features. These limitations hinder the creation of a truly holistic item representation. To overcome this, we propose CEMG, a novel Collaborative-Enhaned Multimodal Generative Recommendation framework. Our approach features a Multimodal Fusion Layer that dynamically integrates visual and textual features under the guidance of collaborative signals. Subsequently, a Unified Modality Tokenization stage employs a Residual Quantization VAE (RQ-VAE) to convert this fused representation into discrete semantic codes. Finally, in the End-to-End Generative Recommendation stage, a large language model is fine-tuned to autoregressively generate these item codes. Extensive experiments demonstrate that CEMG significantly outperforms state-of-the-art baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2512_21543
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CEMG: Collaborative-Enhanced Multimodal Generative Recommendation
Lin, Yuzhen
Chen, Hongyi
Chen, Xuanjing
Wang, Shaowen
Xu, Ivonne
Jiang, Dongming
Information Retrieval
Generative recommendation models often struggle with two key challenges: (1) the superficial integration of collaborative signals, and (2) the decoupled fusion of multimodal features. These limitations hinder the creation of a truly holistic item representation. To overcome this, we propose CEMG, a novel Collaborative-Enhaned Multimodal Generative Recommendation framework. Our approach features a Multimodal Fusion Layer that dynamically integrates visual and textual features under the guidance of collaborative signals. Subsequently, a Unified Modality Tokenization stage employs a Residual Quantization VAE (RQ-VAE) to convert this fused representation into discrete semantic codes. Finally, in the End-to-End Generative Recommendation stage, a large language model is fine-tuned to autoregressively generate these item codes. Extensive experiments demonstrate that CEMG significantly outperforms state-of-the-art baselines.
title CEMG: Collaborative-Enhanced Multimodal Generative Recommendation
topic Information Retrieval
url https://arxiv.org/abs/2512.21543