本文将围绕近四年来发表在顶级会议（NeurIPS、SIGIR、KDD、EMNLP、WSDM、COLING）上的 9 篇代表性工作介绍对话推荐系统领域的问题定义、处理方案和结构演化。另外笔者整理了近年来该领域 68 篇研究工作及其代码和数据集附在文末，分享给读者学习参考。

本文目录如下：

  • 引言
    • 对话推荐概述
    • 整体架构
    • 发展现状
  • 基于属性的对话推荐
    • 问题定义
    • 单轮场景 - CRM
    • 多轮场景 - EAR
    • 搜索剪枝 - CPR
    • 统一架构 - UNICORN
  • 生成式对话推荐
    • 问题定义
    • 降噪协同 - ReDial
    • 知识增强 - KBRD
    • 语义融合 - KGSF
    • 话题引导 - TGReDial
    • 可控生成 - NTRD
  • 总结
    • 未来展望
    • 论文总结

参考文献

[1]. Advances and Challenges in Conversational Recommender Systems: A Survey.

[2]. Conversational Recommender System.

[3]. Estimation–Action–Reflection: Towards Deep Interaction Between Conversational and Recommender Systems.

[4]. Interactive Path Reasoning on Graph for Conversational Recommendation.

[5]. Unified Conversational Recommendation Policy Learning via Graph-based Reinforcement Learning.

[6]. Towards Deep Conversational Recommendations.

[7]. Towards Knowledge-Based Recommender Dialog System.

[8]. Improving Conversational Recommender Systems via Knowledge Graph based Semantic Fusion.

[9]. Towards Topic-Guided Conversational Recommender System.

[10]. Learning Neural Templates for Recommender Dialogue System.

[11]. CRSLab: An Open-Source Toolkit for Building Conversational Recommender System.

内容中包含的图片若涉及版权问题，请及时与我们联系删除