Yi Liu

Yi Liu

Auto-generated profile, based on publicly available information

Yi Liu

Institute of Artificial Intelligence, Xiamen University, China

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 October 2025
Verified
Conference Paper
Open access
GXGuipeng Xv
Guipeng Xv
XLXinyu Li
Xinyu Li
YLYi Liu
Yi Liu
CLChen Lin
Chen Lin
XWXiaoli Wang
Xiaoli Wang

30 September 2026 in ACM Transactions on Information Systems
Verified
Journal Article
Open access
GXGuipeng Xv
Guipeng Xv
YLYi Liu
Yi Liu
XLXinyu Li
Xinyu Li
ZHZhenhua Huang
Zhenhua Huang
CLChen Lin
Chen Lin
Multi-modal recommender system (MRS) has emerged as a key information retrieval technology, widely adopted to enhance various web platforms. However, three interrelated challenges remain insufficiently explored: (1) noisy multi-modal content, (2) noisy user feedback, and (3) misalignment between multi-modal content and user feedback. Previous works have either overlooked these challenges or proposed a complex solution. To tackle these challenges in a lightweight way, we propose L ightweight D enoising and A ligning for M ulti-modal R ecommender S ystem (LDA-MRS). During graph construction, LDA-MRS only constructs a single item-item graph based on consistent cross-modal similarity and dynamic user behavior, effectively reducing noise in multi-modal content. We provide a Lightweight Static Strategy and an Accurate Dynamic Strategy for fusing the graphs. During supervised learning, LDA-MRS leverages multi-modal content to estimate the probability of pairwise observed feedback and introduces a Lightweight Denoising BPR Loss to effectively denoise user feedback. During alignment, LDA-MRS uses Lightweight Alignment guided by User preference to improve task-specific alignment and Lightweight Alignment guided by graded Item relations to achieve finer-grained alignment. LDA-MRS is a lightweight and model-agnostic framework. Experiments on different datasets, backbones, and noisy situations show that LDA-MRS consistently delivers significant performance improvements, highlighting the robustness and effectiveness of LDA-MRS in different conditions.