DeGRe: Dense-supervised Generative Reranking for Recommendation
This work proposes DeGRe (Dense-supervised Generative Reranking), a generative reranking framework that bridges the gap between offline exploration and online efficiency through dense supervision, and demonstrates that DeGRe outperforms baseline models on public benchmarks and industrial datasets.
Chaotian Song, Jing-Yao Zhang, Chenghao Chen et al.
· Proceedings of the 32nd ACM... · 1 citation