MVP: Winning Solution to SMP Challenge 2025 Video Track

2025-07-01

AI Summary

This study addresses the task of social media video popularity prediction by proposing a Multimodal Video Prediction (MVP) framework. Methodologically, MVP integrates deep visual features extracted from pretrained video models with user metadata and contextual information to construct robust multimodal representations; it further employs systematic preprocessing—including logarithmic transformation and outlier filtering—alongside a gradient-boosting regressor to effectively model cross-modal interactions and capture complex nonlinear patterns. Key contributions include: (i) an end-to-end multimodal fusion architecture that explicitly models synergistic effects among video content, user behavior, and situational context; and (ii) a robust, task-specific preprocessing paradigm tailored for popularity forecasting. Evaluated on the official SMP Challenge 2025 video track benchmark, MVP achieved first place, demonstrating substantial improvements in trend detection accuracy and recommendation reliability.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Social media platforms serve as central hubs for content dissemination, opinion expression, and public engagement across diverse modalities. Accurately predicting the popularity of social media videos enables valuable applications in content recommendation, trend detection, and audience engagement. In this paper, we present Multimodal Video Predictor (MVP), our winning solution to the Video Track of the SMP Challenge 2025. MVP constructs expressive post representations by integrating deep video features extracted from pretrained models with user metadata and contextual information. The framework applies systematic preprocessing techniques, including log-transformations and outlier removal, to improve model robustness. A gradient-boosted regression model is trained to capture complex patterns across modalities. Our approach ranked first in the official evaluation of the Video Track, demonstrating its effectiveness and reliability for multimodal video popularity prediction on social platforms. The source code is available at https://anonymous.4open.science/r/SMPDVideo.

Problem

• Predicting popularity of social media videos accurately
• Integrating multimodal features for video representation
• Improving robustness in popularity prediction models

Innovation

• Integrates deep video features with metadata
• Applies systematic preprocessing for robustness
• Uses gradient-boosted regression for multimodal patterns