Mutual Wasserstein Discrepancy Minimization for Sequential Recommendation
Mutual Wasserstein Discrepancy Minimization for Sequential Recommendation
复制标题
DOI：
10.1145/3543507.3583529
复制
发表时间：
2023-01
期刊：
影响因子：
--
通讯作者：
Ziwei Fan;Zhiwei Liu;Hao Peng;Philip S. Yu
中科院分区：
文献类型：
--
作者：
Ziwei Fan;Zhiwei Liu;Hao Peng;Philip S. Yu
自监督顺序推荐通过设计良好的数据增强来最大化互信息，从而显著提高推荐性能。然而，互信息估计是基于Kullback-Leibler散度的计算，存在估计不对称、样本量的指数需求和训练不稳定性等局限性。此外，现有的数据扩充大多是随机的，可能会因随机修改而破坏顺序相关性。这两个问题促使我们研究一种替代的鲁棒互信息测量方法，这种方法能够建模不确定性并减轻KL散度的局限性。为此，我们提出了一种基于互瓦瑟斯坦差异最小化（MStein）的序列推荐自监督学习框架。我们提出了Wasserstein差异度量来度量增广序列之间的互信息。Wasserstein差异测量建立在2-Wasserstein距离的基础上，它更稳健，在小批量中更有效，并且能够模拟随机增强过程的不确定性。我们还提出了一种新的基于Wasserstein差异测量的对比学习损失。在四个基准数据集上的大量实验证明了MStein在基线上的有效性。更多的定量分析显示了对扰动的鲁棒性和批次大小的训练效率。最后，改进分析指出了具有显著不确定性的流行用户/项目的更好表示。源代码在https://github.com/zfan20/MStein。
Self-supervised sequential recommendation significantly improves recommendation performance by maximizing mutual information with well-designed data augmentations. However, the mutual information estimation is based on the calculation of Kullback–Leibler divergence with several limitations, including asymmetrical estimation, the exponential need of the sample size, and training instability. Also, existing data augmentations are mostly stochastic and can potentially break sequential correlations with random modifications. These two issues motivate us to investigate an alternative robust mutual information measurement capable of modeling uncertainty and alleviating KL divergence’s limitations. To this end, we propose a novel self-supervised learning framework based on the Mutual WasserStein discrepancy minimization (MStein) for the sequential recommendation. We propose the Wasserstein Discrepancy Measurement to measure the mutual information between augmented sequences. Wasserstein Discrepancy Measurement builds upon the 2-Wasserstein distance, which is more robust, more efficient in small batch sizes, and able to model the uncertainty of stochastic augmentation processes. We also propose a novel contrastive learning loss based on Wasserstein Discrepancy Measurement. Extensive experiments on four benchmark datasets demonstrate the effectiveness of MStein over baselines. More quantitative analyses show the robustness against perturbations and training efficiency in batch size. Finally, improvements analysis indicates better representations of popular users/items with significant uncertainty. The source code is in https://github.com/zfan20/MStein.