A Vlogger-augmented Graph Neural Network Model for Micro-video Recommendation

TL;DR

This paper addresses micro-video recommendation by incorporating vloggers as a key relational signal. It introduces VA-GNN, which constructs a tripartite user–video–vlogger graph and learns two-view embeddings (video-view and vlogger-view) with cross-view contrastive learning, then adaptively fuses these views for prediction. Through multi-task training and extensive experiments on two real-world datasets, VA-GNN consistently outperforms strong GNN-based and contrastive baselines, demonstrating the value of vlogger information and cross-view alignment. The approach offers a scalable, signal-rich framework for improving recommendation quality in vlogger-driven micro-video ecosystems, with practical implications for streaming platforms and content discovery.

Abstract

Existing micro-video recommendation models exploit the interactions between users and micro-videos and/or multi-modal information of micro-videos to predict the next micro-video a user will watch, ignoring the information related to vloggers, i.e., the producers of micro-videos. However, in micro-video scenarios, vloggers play a significant role in user-video interactions, since vloggers generally focus on specific topics and users tend to follow the vloggers they are interested in. Therefore, in the paper, we propose a vlogger-augmented graph neural network model VA-GNN, which takes the effect of vloggers into consideration. Specifically, we construct a tripartite graph with users, micro-videos, and vloggers as nodes, capturing user preferences from different views, i.e., the video-view and the vlogger-view. Moreover, we conduct cross-view contrastive learning to keep the consistency between node embeddings from the two different views. Besides, when predicting the next user-video interaction, we adaptively combine the user preferences for a video itself and its vlogger. We conduct extensive experiments on two real-world datasets. The experimental results show that VA-GNN outperforms multiple existing GNN-based recommendation models.

Figures (4)

• Figure 1: Architecture of model VA-GNN.
• Figure 2: Sensitivity of vlogger loss weight $\lambda_1$ on two datasets.
• Figure 3: Sensitivity of contrastive loss weight $\lambda_2$ on two datasets.
• Figure 4: Sensitivity of the temperature $\tau$ on two datasets.