OpenP5: An Open-Source Platform for Developing, Training, and Evaluating Large Language Model-based Recommender Systems
OpenP5 is an open-source platform designed to facilitate the development, training, and evaluation of Large Language Model (LLM)-based generative recommender systems for research purposes.
Enhancing Top-k Recommendations with Instruction-Tuned Large Language Models
Instruction tuning large language models as rankers can significantly improve the performance of top-k recommendations by leveraging high-quality training data, position-aware prompts, and the integration of signals from conventional recommender systems.
Leveraging Large Language Models and Contrastive Learning to Enhance Content-based Recommendation
The core message of this paper is to propose a novel method called LoID that leverages large language models and contrastive learning to effectively extract and align semantic information from user-item contents, enabling enhanced content-based recommendation performance across multiple domains.
Review-Based Cross-Domain Recommendation via Hyperbolic Embedding and Hierarchy-Aware Domain Disentanglement
A novel cross-domain recommendation algorithm called HEAD (Hyperbolic Embedding and Hierarchy-Aware Domain Disentanglement) that enhances previous review-based domain disentanglement by incorporating hyperbolic geometry and preserving the hierarchical structure during knowledge transfer.
© 2024 by Linnk AI