Chuan Yuan

PRIME reveals that preserving the original Dense path while adding conditional residuals can significantly enhance CTR model performance without sacrificing stability.

SmartGR achieves an 8.6% boost in recommendation performance while slashing inference time by over 2.3 times, tackling unique challenges in generative recommendation systems.

IV-CoT reveals that separating structural planning from appearance rendering can dramatically enhance the fidelity of text-to-image generation.

Generative recommendation gets a boost: TrieRec leverages the inherent hierarchical structure of item tokens to significantly improve next-item prediction accuracy.