Skip to main content
QUICK REVIEW

[论文解读] Causal Disentangled Recommendation Against User Preference Shifts

Wenjie Wang, Xinyu Lin|arXiv (Cornell University)|Mar 28, 2023
Recommender Systems and Techniques被引用 6
一句话总结

本文提出因果解耦推荐(CDR)框架,通过时间变分自编码器建模用户偏好在时间序列环境中的变化,并利用可学习矩阵解耦稀疏的、按类别划分的偏好影响。CDR 通过联合捕捉偏移动态与稀疏性,提升了分布外(OOD)设置下的泛化能力,在三个数据集上显著提升了推荐准确率与鲁棒性。

ABSTRACT

Recommender systems easily face the issue of user preference shifts. User representations will become out-of-date and lead to inappropriate recommendations if user preference has shifted over time. To solve the issue, existing work focuses on learning robust representations or predicting the shifting pattern. There lacks a comprehensive view to discover the underlying reasons for user preference shifts. To understand the preference shift, we abstract a causal graph to describe the generation procedure of user interaction sequences. Assuming user preference is stable within a short period, we abstract the interaction sequence as a set of chronological environments. From the causal graph, we find that the changes of some unobserved factors (e.g., becoming pregnant) cause preference shifts between environments. Besides, the fine-grained user preference over categories sparsely affects the interactions with different items. Inspired by the causal graph, our key considerations to handle preference shifts lie in modeling the interaction generation procedure by: 1) capturing the preference shifts across environments for accurate preference prediction, and 2) disentangling the sparse influence from user preference to interactions for accurate effect estimation of preference. To this end, we propose a Causal Disentangled Recommendation (CDR) framework, which captures preference shifts via a temporal variational autoencoder and learns the sparse influence from multiple environments. Specifically, an encoder is adopted to infer the unobserved factors from user interactions while a decoder is to model the interaction generation process. Besides, we introduce two learnable matrices to disentangle the sparse influence from user preference to interactions. Lastly, we devise a multi-objective loss to optimize CDR. Extensive experiments on three datasets show the superiority of CDR.

研究动机与目标

  • 为解决推荐系统中用户偏好偏移带来的挑战,此类偏移会降低模型在分布外(OOD)环境下的性能。
  • 通过建模未观测因素(如人生事件)及其在时间有序交互环境中的影响,提供对偏好偏移的全面因果理解。
  • 通过解耦用户偏好在类别层面的稀疏影响,提升交互预测的泛化能力。
  • 构建一个统一框架,通过多环境学习与因果结构建模,同时捕捉不变偏好与变化偏好。

提出的方法

  • 该框架将用户交互建模为一系列时间有序的环境,每个环境代表一个相对稳定的用户偏好时间段。
  • 采用时间变分自编码器(TVAE)从交互中推断未观测因素(如怀孕)并建模环境间偏好偏移。
  • 引入两个可学习矩阵,以解耦类别层面用户偏好对物品交互的稀疏影响,实现细粒度效应估计。
  • 设计多目标损失函数,结合方差与稀疏性正则化,联合优化偏好偏移建模与稀疏影响学习。
  • 模型支持在训练与推理阶段灵活调整环境数量,适应不同数据分布。

实验结果

研究问题

  • RQ1驱动用户偏好在不同时间有序交互环境间发生偏移的潜在因果因素是什么?
  • RQ2如何在区分稳定与变化成分的前提下,建模用户偏好从一个环境到另一个环境的转变?
  • RQ3解耦类别层面偏好的稀疏影响在多大程度上能提升分布外(OOD)设置下的交互预测准确率?
  • RQ4多环境、因果感知框架是否能在偏好偏移场景下,优于现有鲁棒或序列推荐模型,展现出更好的泛化能力?

主要发现

  • CDR 在用户偏好偏移场景下显著提升了推荐性能,在三个真实世界数据集上,其准确率与泛化能力均优于最先进基线模型。
  • 消融实验确认,用于偏移建模的时间VAE与解耦稀疏影响矩阵均对性能提升有显著贡献。
  • 结合方差与稀疏性正则化的多目标损失提升了模型稳定性与泛化能力,尤其在分布外评估场景中表现突出。
  • 该框架在不同环境数量下均表现出鲁棒性,展示了在真实世界部署中对不同数据划分策略的灵活性。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。