no code implementations • 28 Feb 2023 • Shenzheng Zhang, Qi Tan, Xinzhi Zheng, Yi Ren, Xu Zhao
The gap between the randomly initialized item ID embedding and the well-trained warm item ID embedding makes the cold items hard to suit the recommendation system, which is trained on the data of historical warm items.
1 code implementation • 24 Feb 2023 • Yi Ren, Xiao Han, Xu Zhao, Shenzheng Zhang, Yan Zhang
Therefore, the ranking stage is still essential for most applications to provide high-quality candidate set for the re-ranking stage.
1 code implementation • 27 May 2022 • Xu Zhao, Yi Ren, Ying Du, Shenzheng Zhang, Nian Wang
This paper attempts to tackle the item cold-start problem by generating enhanced warmed-up ID embeddings for cold items with historical data and limited interaction records.