arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.09445cs.IR
收藏

面向多模态推荐的个性化参数高效微调

Personalized Parameter-Efficient Fine-Tuning of Foundation Models for Multimodal Recommendation

KAIST(韩国科学技术院)

Sunwoo Kim, Hyunjin Hwang, Kijung Shin

AI总结:

本文提出PerPEFT,一种面向多模态推荐的个性化参数高效微调策略,通过按兴趣分组并为每个组分配独立模块,提升推荐效果。

摘要:

近年来,大量研究将多模态物品元数据整合到推荐系统中,通常通过使用预训练的多模态基础模型来编码此类数据。由于这些模型并非最初为推荐任务训练,最近的研究通过参数高效微调(PEFT)来高效地适应它们。然而,即使使用PEFT,多模态基础模型的物品嵌入仍然缺乏用户感知:物品嵌入未根据用户兴趣进行条件化,尽管具有不同兴趣的用户会关注不同的物品方面。为了解决这一限制,我们提出了PerPEFT,一种面向多模态推荐的个性化PEFT策略。具体而言,PerPEFT按兴趣将用户分组,并为每个组分配一个独特的PEFT模块,使每个模块能够捕捉最能预测该组购买决策的细粒度物品方面。我们进一步引入了一种专门的训练技术,以加强这种用户-组条件化。值得注意的是,PerPEFT是PEFT无关的,可以与任何适用于多模态基础模型的PEFT方法配对。通过广泛的实验,我们证明了(1)PerPEFT在NDCG@20上比最强基线高出高达15.3%(2)在各种PEFT变体中均能提供一致的提升。值得注意的是,即使有个性化,PEFT仍保持轻量级,仅增加基础模型参数量的1.3%。我们提供了代码和数据集:https://github.com/kswoo97/PerPEFT。

英文摘要:

In recent years, substantial research has integrated multimodal item metadata into recommender systems, often by using pre-trained multimodal foundation models to encode such data. Since these models are not originally trained for recommendation tasks, recent works efficiently adapt them via parameter-efficient fine-tuning (PEFT). However, even with PEFT, item embeddings from multimodal foundation models remain user-blind: item embeddings are not conditioned on user interests, despite the fact that users with diverse interests attend to different item aspects. To address this limitation, we propose PerPEFT, a personalized PEFT strategy for multimodal recommendation. Specifically, PerPEFT groups users by interest and assigns a distinct PEFT module to each group, enabling each module to capture the fine-grained item aspects most predictive of that group`s purchase decisions. We further introduce a specialized training technique that strengthens this user-group conditioning. Notably, PerPEFT is PEFT-agnostic and can be paired with any PEFT method applicable to multimodal foundation models. Through extensive experiments, we show that (1) PerPEFT outperforms the strongest baseline by up to 15.3% (NDCG@20) and (2) delivers consistent gains across diverse PEFT variants. It is noteworthy that, even with personalization, PEFT remains lightweight, adding only 1.3% of the parameter count of the foundation model. We provide our code and datasets at https://github.com/kswoo97/PerPEFT.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑