M3TR: Temporal Retrieval Enhanced Multi-Modal Micro-video Popularity Prediction

Accurately predicting the popularity of micro-videos is a critical but challenging task, characterized by volatile, `rollercoaster-like' engagement dynamics. Existing methods often fail to capture these complex temporal patterns, leading to inaccurate long-term forecasts. This failure stems from two fundamental limitations: ding172 a superficial understanding of user feedback dynamics, which overlooks the mutually exciting and decaying nature of interactions such as likes, comments, and shares; and~ding173 retrieval mechanisms that rely solely on static content similarity, ignoring the crucial patterns of how a video's popularity evolves over time. To address these limitations, we propose M3TR, a Temporal Retrieval enhanced Multi-Modal framework that uniquely synergizes fine-grained temporal modeling with a novel temporal-aware retrieval process for Micro-video popularity prediction. At its core, M3TR introduces a Mamba-Hawkes Process (MHP) module to explicitly model user feedback as a sequence of self-exciting events, capturing the intricate, long-range dependencies within user interactions (for limitation ding172). This rich temporal representation then powers a temporal-aware retrieval engine that identifies historically relevant videos based on a combined similarity of both their multi-modal content (visual, audio, text) and their popularity trajectories (for limitation ding173). By augmenting the target video's features with this retrieved knowledge, M3TR achieves a comprehensive understanding of prediction. Extensive experiments on two real-world datasets demonstrate the superiority of our framework. M3TR achieves state-of-the-art performance, outperforming previous methods by up to 19.3% in nMSE and showing significant gains in addressing long-term prediction challenges.