| Autor: |
Zhou, Yujia, Yao, Jing, Dou, Zhicheng, Tu, Yiteng, Wu, Ledell, Chua, Tat-Seng, Wen, Ji-Rong |
| Zdroj: |
ACM Transactions on Information Systems; Nov2024, Vol. 42 Issue 6, p1-25, 25p |
| Abstrakt: |
The article discusses the ranking-oriented generative retrieval (ROGER) model that integrates relevance feedback from dense retrieval to improve ranking performance. Topics include the limitations of existing generative retrieval models, the complementary nature of generative and dense retrieval, and ROGER's experimental success in enhancing recall on datasets. It suggests to explore reinforcement learning and large language models to further optimize document retrieval systems. |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
