Repeated Padding+：一种用于序列推荐的简单而有效的数据增强插件
Repeated Padding+: Simple yet Effective Data Augmentation Plugin for Sequential Recommendation
AI总结：
提出一种名为RepPad+的即插即用序列级数据增强方法，通过将原始交互序列填充至闲置空间，无需额外参数即可显著提升多种序列推荐模型的性能。
摘要：
序列推荐旨在根据用户的历史交互为其提供个性化建议。在训练序列模型时，填充是一种被广泛采用的技术，主要原因有两点：1）绝大多数模型只能处理固定长度的序列；2）基于批处理的训练需要确保每个批次中的序列具有相同的长度。特殊值0通常被用作填充内容，它不包含实际信息，并在模型计算中被忽略。这种常识性的填充策略引出了一个此前从未被探索过的问题：我们能否通过填充其他内容来充分利用这一闲置的输入空间，从而进一步提升模型性能和训练效率？在这项工作中，我们提出了一种简单而有效的填充方法，称为Repeated Padding+（RepPad+）。具体而言，我们将原始交互序列作为填充内容，并在模型训练期间将其填充到填充位置。该操作可以执行有限次数，或重复执行直到输入序列的长度达到最大限制。对于那些无法用完整原始数据填满的序列，我们从Sliding Windows策略中汲取灵感，截取连续的子序列来填充闲置空间。我们的RepPad+可以被视为一种序列级数据增强策略。与大多数现有工作不同，我们的方法不包含任何可训练参数或超参数，是一种即插即用的数据增强操作。在各类序列模型和七个真实世界数据集上的广泛实验证明了我们方法的有效性和高效性。在GRU4Rec上平均推荐性能提升高达84.11%，在SASRec上提升35.34%。我们还从多个视角对RepPad+有效的原因提供了深入的分析和解释。
英文摘要：
Sequential recommendation aims to provide users with personalized suggestions based on their historical interactions. When training sequential models, padding is a widely adopted technique for two main reasons: 1) The vast majority of models can only handle fixed-length sequences; 2) Batching-based training needs to ensure that the sequences in each batch have the same length. The special value \emph{0} is usually used as the padding content, which does not contain the actual information and is ignored in the model calculations. This common-sense padding strategy leads us to a problem that has never been explored before: Can we fully utilize this idle input space by padding other content to further improve model performance and training efficiency? In this work, we propose a simple yet effective padding method called Repeated Padding+ (RepPad+). Specifically, we use the original interaction sequences as the padding content and fill it to the padding positions during model training. This operation can be performed a finite number of times or repeated until the input sequences' length reaches the maximum limit. For those sequences that can not pad full original data, we draw inspiration from the Sliding Windows strategy and intercept consecutive subsequences to fill in the idle space. Our RepPad+ can be viewed as a sequence-level data augmentation strategy. Unlike most existing works, our method contains no trainable parameters or hyperparameters and is a plug-and-play data augmentation operation. Extensive experiments on various categories of sequential models and seven real-world datasets demonstrate the effectiveness and efficiency of our approach. The average recommendation performance improvement is up to 84.11% on GRU4Rec and 35.34% on SASRec. We also provide in-depth analysis and explanation of what makes RepPad+ effective from multiple perspectives.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
