[论文解读] Personalized Prompt Learning for Explainable Recommendation

Lei Li, Yongfeng Zhang+1 more|arXiv (Cornell University)|Feb 15, 2022

Recommender Systems and Techniques被引用 29

一句话总结

PEPLER 引入个性化提示学习，通过将用户/项目ID视为提示来生成可解释的推荐，使用离散提示（特征）或连续提示（ID向量），并提供两种训练策略以弥合与预训练变换器之间的差距

ABSTRACT

Providing user-understandable explanations to justify recommendations could help users better understand the recommended items, increase the system's ease of use, and gain users' trust. A typical approach to realize it is natural language generation. However, previous works mostly adopt recurrent neural networks to meet the ends, leaving the potentially more effective pre-trained Transformer models under-explored. In fact, user and item IDs, as important identifiers in recommender systems, are inherently in different semantic space as words that pre-trained models were already trained on. Thus, how to effectively fuse IDs into such models becomes a critical issue. Inspired by recent advancement in prompt learning, we come up with two solutions: find alternative words to represent IDs (called discrete prompt learning), and directly input ID vectors to a pre-trained model (termed continuous prompt learning). In the latter case, ID vectors are randomly initialized but the model is trained in advance on large corpora, so they are actually in different learning stages. To bridge the gap, we further propose two training strategies: sequential tuning and recommendation as regularization. Extensive experiments show that our continuous prompt learning approach equipped with the training strategies consistently outperforms strong baselines on three datasets of explainable recommendation.

研究动机与目标

• 激发对推荐系统中用户可理解解释的需求，以提升可用性和信任度。
• 研究如何将用户/项目ID与预训练语言模型融合，以生成解释。
• 提出两种基于提示的方法（离散和连续），不使用传统序列模型来生成解释。
• 开发训练策略，使随机初始化的提示与固定的预训练模型对齐，以提升可解释性。

提出的方法

• 提出离散提示学习（PEPLER-D），通过将项特征用作提示，在预训练变换器内引导解释生成。
• 提出连续提示学习（PEPLER-C），通过将ID嵌入为可训练的提示向量，并将它们与解释一起输入。
• 引入负对数似然目标，以在条件于提示的情况下训练模型进行词预测。
• 为连续提示开发两种训练策略：序列微调（冻结语言模型后进行联合微调）和以推荐作为正则化（使用评分预测来提升解释质量）。
• 将用户和项目ID表示为特征（离散提示）或学习得到的嵌入（连续提示），并通过贪婪解码生成解释。
• 确保与任何基础推荐模型的兼容性，通过对(u,i)或带特征的(u,i)进行解释条件化。

Figure 1. A review example from Yelp. The user and the restaurant are omitted for privacy protection.

实验结果

研究问题

• RQ1 个性化提示（离散或连续）是否能利用预训练语言模型实现有效的可解释推荐？
• RQ2 如何弥合随机初始化的连续提示与固定的预训练变换器之间的学习差距？
• RQ3 如序列微调与推荐作为正则化等训练策略是否提高了解释质量与可解释性？
• RQ4 离散（基于特征）提示与连续（ID向量）提示在生成解释方面有何差异？
• RQ5 该方法是否适用于不同的底层推荐模型（如 MF、MLP）和数据集？

主要发现

• 在三个数据集上，结合所提出的训练策略的连续提示学习在文本质量和可解释性方面优于强基线。
• PEPLER-D 在预训练模型内利用信息丰富的项特征作为提示来实现有效的解释生成。
• 序列微调在联合训练之前帮助将连续提示与预训练模型对齐，降低优化问题。
• 将推荐作为正则化通过将评分预测信号纳入学习过程来提升解释质量。
• 在可解释推荐任务中，PEPLER 相较于基于 RNN、标准 Transformer 或非预训练方法的效果更佳。

Figure 2. A comparison between left-to-right unidirectional attention masking (left) and bidirectional attention masking (right).