Researchers have introduced M3TR, a novel framework designed to improve the prediction of micro-video popularity. This system addresses limitations in current methods by incorporating a Mamba-Hawkes Process module to better model user engagement dynamics and a temporal-aware retrieval engine that considers both content and popularity evolution. Experiments show M3TR significantly outperforms existing approaches, achieving up to a 19.3% improvement in nMSE on real-world datasets. AI
IMPACT This framework could lead to more accurate content recommendation and trend forecasting in the social media landscape.
RANK_REASON The cluster contains an academic paper detailing a new framework for a specific AI task. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
