Back to Search
Start Over
MCRPL: A Pretrain, Prompt, and Fine-tune Paradigm for Non-overlapping Many-to-one Cross-domain Recommendation.
- Source :
-
ACM Transactions on Information Systems . Jul2024, Vol. 42 Issue 4, p1-24. 24p. - Publication Year :
- 2024
-
Abstract
- The article introduces MCRPL, a paradigm combining pretraining, prompt, and fine-tuning, to address the challenge of non-overlapping many-to-one cross-domain recommendation. Topics include the integration of domain-agnostic and domain-dependent prompts, updating prompts for knowledge transfer, and experimental validation against recent baselines, demonstrating superior performance.
Details
- Language :
- English
- ISSN :
- 10468188
- Volume :
- 42
- Issue :
- 4
- Database :
- Academic Search Index
- Journal :
- ACM Transactions on Information Systems
- Publication Type :
- Academic Journal
- Accession number :
- 177224588
- Full Text :
- https://doi.org/10.1145/3641860