Xichen Ding
Research Subjects
Research Interests
Xichen Ding has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Guannan Zhang
Alibaba Group, Hangzhou, Zhejiang, China
1 shared publication - Kaiming Shen
Ant Group, Beijing, China
1 shared publication - Yehui Su
Ant Group, Beijing, China
1 shared publication - Yuqi Gong
Ant Group, Beijing, China
1 shared publication - Zhongyi Liu
Ant Group, Hangzhou, China
1 shared publication
Latest Publications
- Open Access
- 21 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 4595-4601
An Unified Search and Recommendation Foundation Model for Cold-Start Scenario
In modern commercial search engines and recommendation systems, data from multiple domains is available to jointly train the multi-domain model. Traditional methods train multi-domain models in the multi-task setting, with shared parameters to learn the similarity of multiple tasks, and task-specific parameters to learn the divergence of features, labels, and sample distributions of individual tasks. With the development of large language models, LLM can extract global domain-invariant text features that serve both search and recommendation tasks. We propose a novel framework called S&R Multi-Domain Foundation, which uses LLM to extract domain invariant features, and Aspect Gating Fusion to merge the ID feature, domain invariant text features and task-specific heterogeneous sparse features to obtain the representations of query and item. Additionally, samples from multiple search and recommendation scenarios are trained jointly with Domain Adaptive Multi-Task module to obtain the multi-domain foundation model. We apply the S&R Multi-Domain foundation model to cold start scenarios in the pretrain-finetune manner, which achieves better performance than other SOTA transfer learning methods. The S&R Multi-Domain Foundation model has been successfully deployed in Alipay Mobile Application's online services, such as content query recommendation and service card recommendation, etc.
- 21 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 4595-4601
Open AccessAn Unified Search and Recommendation Foundation Model for Cold-Start Scenario
In modern commercial search engines and recommendation systems, data from multiple domains is available to jointly train the multi-domain model. Traditional methods train multi-domain models in the multi-task setting, with shared parameters to learn the similarity of multiple tasks, and task-specific parameters to learn the divergence of features, labels, and sample distributions of individual tasks. With the development of large language models, LLM can extract global domain-invariant text features that serve both search and recommendation tasks. We propose a novel framework called S&R Multi-Domain Foundation, which uses LLM to extract domain invariant features, and Aspect Gating Fusion to merge the ID feature, domain invariant text features and task-specific heterogeneous sparse features to obtain the representations of query and item. Additionally, samples from multiple search and recommendation scenarios are trained jointly with Domain Adaptive Multi-Task module to obtain the multi-domain foundation model. We apply the S&R Multi-Domain foundation model to cold start scenarios in the pretrain-finetune manner, which achieves better performance than other SOTA transfer learning methods. The S&R Multi-Domain Foundation model has been successfully deployed in Alipay Mobile Application's online services, such as content query recommendation and service card recommendation, etc.