Learning Decomposed Contextual Token Representations from Pretrained and Collaborative Signals for Generative Recommendation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Liu, Yifan, Liu, Yaokun, Li, Zelin, Yue, Zhenrui, Lee, Gyuseok, Yao, Ruichen, Zhang, Yang, Wang, Dong
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913081645334528
author Liu, Yifan
Liu, Yaokun
Li, Zelin
Yue, Zhenrui
Lee, Gyuseok
Yao, Ruichen
Zhang, Yang
Wang, Dong
author_facet Liu, Yifan
Liu, Yaokun
Li, Zelin
Yue, Zhenrui
Lee, Gyuseok
Yao, Ruichen
Zhang, Yang
Wang, Dong
contents Recent advances in generative recommenders adopt a two-stage paradigm: items are first tokenized into semantic IDs using a pretrained tokenizer, and then large language models (LLMs) are trained to generate the next item via sequence-to-sequence modeling. However, these two stages are optimized for different objectives: semantic reconstruction during tokenizer pretraining versus user interaction modeling during recommender training. This objective misalignment leads to two key limitations: (i) suboptimal static tokenization, where fixed token assignments fail to reflect diverse usage contexts; and (ii) discarded pretrained semantics, where pretrained knowledge - typically from language model embeddings - is overwritten during recommender training on user interactions. To address these limitations, we propose to learn $\underline{DE}$composed $\underline{CO}$ntextual Token $\underline{R}$epresentations (DECOR), a unified framework that preserves pretrained semantics while enhancing the adaptability of token embeddings. DECOR introduces contextualized token composition to refine token embeddings based on user interaction context, and decomposed embedding fusion that integrates pretrained codebook embeddings with newly learned collaborative embeddings. Experiments on three real-world datasets demonstrate that DECOR consistently outperforms state-of-the-art baselines in recommendation performance.
format Preprint
id arxiv_https___arxiv_org_abs_2509_10468
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning Decomposed Contextual Token Representations from Pretrained and Collaborative Signals for Generative Recommendation
Liu, Yifan
Liu, Yaokun
Li, Zelin
Yue, Zhenrui
Lee, Gyuseok
Yao, Ruichen
Zhang, Yang
Wang, Dong
Information Retrieval
Artificial Intelligence
Computation and Language
cs.IR
Recent advances in generative recommenders adopt a two-stage paradigm: items are first tokenized into semantic IDs using a pretrained tokenizer, and then large language models (LLMs) are trained to generate the next item via sequence-to-sequence modeling. However, these two stages are optimized for different objectives: semantic reconstruction during tokenizer pretraining versus user interaction modeling during recommender training. This objective misalignment leads to two key limitations: (i) suboptimal static tokenization, where fixed token assignments fail to reflect diverse usage contexts; and (ii) discarded pretrained semantics, where pretrained knowledge - typically from language model embeddings - is overwritten during recommender training on user interactions. To address these limitations, we propose to learn $\underline{DE}$composed $\underline{CO}$ntextual Token $\underline{R}$epresentations (DECOR), a unified framework that preserves pretrained semantics while enhancing the adaptability of token embeddings. DECOR introduces contextualized token composition to refine token embeddings based on user interaction context, and decomposed embedding fusion that integrates pretrained codebook embeddings with newly learned collaborative embeddings. Experiments on three real-world datasets demonstrate that DECOR consistently outperforms state-of-the-art baselines in recommendation performance.
title Learning Decomposed Contextual Token Representations from Pretrained and Collaborative Signals for Generative Recommendation
topic Information Retrieval
Artificial Intelligence
Computation and Language
cs.IR
url https://arxiv.org/abs/2509.10468