Search papers, labs, and topics across Lattice.
This paper introduces PrismRec, a novel framework for micro-video recommendation that utilizes Spectral Semantic Factorization (SSF) to separate static and dynamic factors from frame-level representations. By employing Context-Calibrated Preference Matching (CPM), the method effectively incorporates user-specific sensitivities to enhance preference formation, treating video content as a primary driver rather than a secondary factor. Experimental results demonstrate that PrismRec outperforms state-of-the-art methods by up to 22.65% while maintaining lower inference costs and peak memory usage.
PrismRec revolutionizes micro-video recommendation by treating video content as a core driver of user preferences, achieving a remarkable 22.65% improvement over existing methods.
Micro-video recommendation aims to infer user preferences from historical interactions and multimodal video content, thereby identifying the next video of interest. However, prevailing methods compress frame sequences into a single holistic representation, entangling the stable visual semantics and the evolving dynamics that jointly shape user preferences. Meanwhile, diffusion- and flow matching-based recommenders condition their generation process solely on coarse behavioral context, leaving its internal temporal structure outside preference formation. We therefore propose PrismRec, a Preference Flow Matching framework with Spectral Factorization for Micro-video Recommendation. Analogous to a prism that disperses white light into its constituent spectrum, PrismRec devises Spectral Semantic Factorization (SSF) to derive complementary static semantic and dynamic factors from frame-level representations via a prior-guided learnable frequency mask in the temporal frequency domain. Then, it proposes Context-Calibrated Preference Matching (CPM) to weigh them with each user's specific sensitivity and inject the calibrated context as a structured condition to steer the matching trajectory toward the target representation, making video content as an intrinsic driver of preference formation rather than auxiliary side information. Experiments on four datasets from two platforms show that PrismRec surpasses the SOTA baseline by up to 22.65%, with the lowest inference cost and peak memory among the compared methods.
