核心概念
Incorporating planning capabilities into recommendation systems enhances long-term engagement.
摘要
The content discusses the importance of incorporating planning capabilities into recommendation systems to improve long-term engagement. It introduces a Bi-level Learnable Large Language Model Planning framework for this purpose, highlighting the macro-learning and micro-learning mechanisms. The framework aims to enhance the planning ability of Large Language Models (LLMs) for long-term recommendations through a hierarchical approach. Extensive experiments validate the framework's superiority in learning to plan for long-term recommendations.
统计
Reinforcement Learning (RL) can learn planning capacity by maximizing cumulative reward.
Large Language Models (LLMs) have powerful planning capabilities through pre-training on diverse textual data.
The proposed Bi-level Learnable LLM Planner framework combines macro-learning and micro-learning.
引用
"Traditional recommendation setting tends to excessively cater to users’ immediate interests and neglect their long-term engagement."
"LLMs have emerged with powerful planning capabilities through pre-training on massive and diverse textual data."