Xiangning Chen
University of California, Los Angeles
47 Papers
99 Citations
Xiangning Chen is an academic researcher from University of California, Los Angeles. The author has contributed to research in topics: Computer science & Recommender system. The author has an hindex of 11, co-authored 25 publications. Previous affiliations of Xiangning Chen include Google & Tsinghua University.
Chat about Author
Papers
Symbolic Discovery of Optimization Algorithms
Xiangning Chen,Chen Liang,Da Huang,Esteban Real,Kaiyuan Wang,Yao Liu,Hieu Quang Pham,Xuanyi Dong,Thang Luong,Cho-Jui Hsieh,Yifeng Lu,Quoc Le +11 more
TL;DR: Lion as mentioned in this paper proposes a simple and effective optimization algorithm, which is more memory efficient than Adam as it only keeps track of the momentum of the sign operation, and it also requires a smaller learning rate than Adam due to the larger norm of the update.
Neural Multi-task Recommendation from Multi-behavior Data
Chen Gao,Xiangnan He,Dahua Gan,Xiangning Chen,Fuli Feng,Yong Li,Tat-Seng Chua,Depeng Jin +7 more
- 08 Apr 2019
TL;DR: In this paper, the authors proposed Neural Multi-Task Recommendation (NMTR) for learning recommender systems from user multi-behavior data, which accounts for the cascading relationship among different types of behaviors and performs a joint optimization based on the multi-task learning framework.
186
•Posted Content
Stabilizing Differentiable Architecture Search via Perturbation-based Regularization
Xiangning Chen,Cho-Jui Hsieh +1 more
TL;DR: This work finds that the precipitous validation loss landscape, which leads to a dramatic performance drop when distilling the final architecture, is an essential factor that causes instability and proposes a perturbation-based regularization - SmoothDARTS (SDARTS), to smooth the loss landscape and improve the generalizability of DARTS-based methods.
166
Cross-domain Recommendation Without Sharing User-relevant Data
Chen Gao,Xiangning Chen,Fuli Feng,Kai Zhao,Xiangnan He,Yong Li,Depeng Jin +6 more
- 13 May 2019
TL;DR: To avoid the leak of user privacy during the data sharing process, a new method named NATR (short for Neural Attentive Transfer Recommendation) is considered, making it easier for two companies to reach a consensus on data sharing since the data to be shared is user-irrelevant and has no explicit semantics.
147
•Posted Content
Neural Multi-Task Recommendation from Multi-Behavior Data
Chen Gao,Xiangnan He,Dahua Gan,Xiangning Chen,Fuli Feng,Yong Li,Tat-Seng Chua,Depeng Jin +7 more
- 21 Sep 2018
TL;DR: This work contributes a new solution named NMTR (short for Neural Multi-Task Recommendation) for learning recommender systems from user multi-behavior data, which significantly outperforms state-of-the-artRecommender systems that are designed to learn from both single- behavior data and multi- Behavior data.
126