Contrastive Learning (CL)-based recommender systems have gained prominence in the context of Heterogeneous Graph (HG) due to their capacity to enhance the consistency of representations across different views. Nonetheless, existing frameworks often neglect the fact that user-item interactions within HG are governed by diverse latent intents (for instance, preferences towards specific brands or the demographic characteristics of item audiences), which are pivotal in capturing fine-grained relations. The exploration of these underlying intents, particularly through the lens of meta-paths in HGs, presents us with two principal challenges: i) How to integrate CL mechanisms with latent intents; ii) How to mitigate the noise associated with these complicated intents.To address these challenges, we propose an innovative framework termed Intent-Guided Heterogeneous Graph Contrastive Learning (IHGCL), which designed to enhance CL-based recommendation by capturing the intents contained within meta-paths. Specifically, the IHGCL framework includes: i) it employs a meta-path-based dual contrastive learning approach to effectively integrate intents into the recommendation, constructing meta-path contrast and view contrast; ii) it uses an bottlenecked autoencoder that combines mask propagation with the information bottleneck principle to significantly reduce noise perturbations introduced by meta-paths. Empirical evaluations conducted across six distinct datasets demonstrate the superior performance of our IHGCL framework relative to conventional baseline methods. Our model implementation is available at https://github.com/wangyu0627/IHGCL.

翻译：基于对比学习（CL）的推荐系统因其能够增强不同视图间表示的一致性，在异质图（HG）背景下日益受到重视。然而，现有框架往往忽略了异质图中用户-物品交互受多种潜在意图（例如对特定品牌的偏好或物品受众的人口统计特征）支配这一事实，而这些意图对于捕获细粒度关系至关重要。探索这些潜在意图，特别是通过异质图中元路径的视角，给我们带来了两个主要挑战：i) 如何将对比学习机制与潜在意图相结合；ii) 如何减轻与这些复杂意图相关的噪声。为应对这些挑战，我们提出了一种名为意图引导的异质图对比学习（IHGCL）的创新框架，该框架旨在通过捕获元路径中包含的意图来增强基于对比学习的推荐。具体而言，IHGCL框架包括：i) 采用基于元路径的双重对比学习方法，通过构建元路径对比和视图对比，将意图有效整合到推荐中；ii) 使用结合掩码传播与信息瓶颈原则的瓶颈自编码器，以显著减少元路径引入的噪声干扰。在六个不同数据集上进行的实证评估表明，我们的IHGCL框架相较于传统基线方法具有更优的性能。我们的模型实现可在 https://github.com/wangyu0627/IHGCL 获取。