[论文解读] Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction
LAIN 引入了面向长度的条件化来 CTR 模型,使用光谱长度编码器、长度条件提示和长度调制注意力,在整体 CTR 和短序列用户表现上都实现了提升,开销极小。
User behavior sequences in modern recommendation systems exhibit significant length heterogeneity, ranging from sparse short-term interactions to rich long-term histories. While longer sequences provide more context, we observe that increasing the maximum input sequence length in existing CTR models paradoxically degrades performance for short-sequence users due to attention polarization and length imbalance in training data. To address this, we propose LAIN(Length-Adaptive Interest Network), a plug-and-play framework that explicitly incorporates sequence length as a conditioning signal to balance long- and short-sequence modeling. LAIN consists of three lightweight components: a Spectral Length Encoder that maps length into continuous representations, Length-Conditioned Prompting that injects global contextual cues into both long- and short-term behavior branches, and Length-Modulated Attention that adaptively adjusts attention sharpness based on sequence length. Extensive experiments on three real-world benchmarks across five strong CTR backbones show that LAIN consistently improves overall performance, achieving up to 1.15% AUC gain and 2.25% log loss reduction. Notably, our method significantly improves accuracy for short-sequence users without sacrificing longsequence effectiveness. Our work offers a general, efficient, and deployable solution to mitigate length-induced bias in sequential recommendation.
研究动机与目标
- 识别当前 CTR 模型中长序列用户与短序列用户之间的性能不平衡。
- 提出一个长度自适应框架,在不同序列长度之间平衡兴趣建模。
- 证明长度条件化在不牺牲长序列性能的情况下提升短序列精度。
- 展示 LAIN 在多种骨干网络和数据集上的鲁棒性与可部署性。
提出的方法
- 引入 Spectral Length Encoder,将原始序列长度映射到可训练傅里叶基的连续嵌入。
- 开发 Length-Conditioned Prompting,在短期和长期行为序列前置长度感知提示。
- 实现 Length-Modulated Attention,基于长度先验对查询/键进行条件化并调整 softmax 温度。
- 通过提示解耦共享参数与长度特定参数,建立长度相关的归纳偏置。
- 提供端到端训练,参数/计算开销微小。
- 在三个真实世界数据集与五个强 CTR 骨干网络上评估 LAIN。
实验结果
研究问题
- RQ1显式对序列长度进行条件化是否能够缓解 CTR 模型中的注意力偏极化和长度信号不足?
- RQ2LAIN 是否在不同骨干与数据集上提升短序列用户 CTR,同时不损害长序列性能?
- RQ3每个 LAIN 组件(SLE、LCP、LMA)及其相互作用的经验影响如何?
- RQ4对于超参数选择和不同序列长度分布,LAIN 的鲁棒性如何?
主要发现
| Model | EBNeRD-small GAUC | EBNeRD-small AUC | EBNeRD-small logloss | KuaiVideo GAUC | KuaiVideo AUC | KuaiVideo logloss | MicroVideo1.7M GAUC | MicroVideo1.7M AUC | MicroVideo1.7M logloss |
|---|---|---|---|---|---|---|---|---|---|
| DIN | 0.7053 | 0.7080 | 0.2695 | 0.6716 | 0.6979 | 0.4498 | 0.7023 | 0.7093 | 0.4196 |
| +LAIN | 0.7096 | 0.7156 | 0.2678 | 0.6742 | 0.7043 | 0.4457 | 0.7019 | 0.7101 | 0.4179 |
| Rel. Gain | 0.61% | 1.07% | -0.65% | 0.40% | 0.93% | -0.92% | -0.05% | 0.12% | -0.40% |
| DIEN | 0.7090 | 0.7099 | 0.2737 | 0.6678 | 0.6977 | 0.4509 | 0.7143 | 0.7185 | 0.4171 |
| +LAIN | 0.7078 | 0.7118 | 0.2676 | 0.6688 | 0.6991 | 0.4476 | 0.7146 | 0.7219 | 0.4161 |
| Rel. Gain | -0.17% | 0.27% | -2.25% | 0.14% | 0.20% | -0.71% | 0.04% | 0.47% | -0.23% |
| SIM | 0.6960 | 0.6992 | 0.2720 | 0.6672 | 0.6875 | 0.4574 | 0.7017 | 0.7095 | 0.4168 |
| +LAIN | 0.6990 | 0.7037 | 0.2706 | 0.6678 | 0.6896 | 0.4566 | 0.7070 | 0.7131 | 0.4151 |
| Rel. Gain | 0.43% | 0.65% | -0.50% | 0.08% | 0.31% | -0.16% | 0.75% | 0.50% | -0.41% |
| SDIM | 0.7039 | 0.7099 | 0.2719 | 0.6729 | 0.6924 | 0.4536 | 0.6984 | 0.7161 | 0.4129 |
| +LAIN | 0.7099 | 0.7123 | 0.2704 | 0.6732 | 0.6949 | 0.4525 | 0.6993 | 0.7163 | 0.4121 |
| Rel. Gain | 0.85% | 0.34% | -0.56% | 0.05% | 0.35% | -0.23% | 0.13% | 0.03% | -0.18% |
| TWIN | 0.6930 | 0.6993 | 0.2718 | 0.6729 | 0.6918 | 0.4530 | 0.7060 | 0.7158 | 0.4164 |
| +LAIN | 0.7012 | 0.7074 | 0.2698 | 0.6748 | 0.6976 | 0.4508 | 0.7093 | 0.7233 | 0.4097 |
| Rel. Gain | 1.19% | 1.15% | -0.73% | 0.29% | 0.84% | -0.48% | 0.47% | 1.05% | -1.63% |
- LAIN 在不同骨干和数据集上持续提升整体 CTR 性能,GAUC 最高提升至 +1.2%,logloss 最多降低至 -1.6%。
- LAIN 显著提升短序列用户(少于 100)的准确性,在 MicroVideo1.7M 的 TWIN 上 AUC 提升至 +1.08%,logloss 降至 -2.17%。
- 注意力极化被 LAIN 缓解,在短、中、长序列下的 Gini 系数显著低于基线。
- 消融研究显示所有组件均有贡献;移除 LMA 会导致最大的性能下降。
- 超参数敏感性分析显示在不同配置下仍能获得鲁棒提升(选定设置:傅里叶维度 64,隐藏层 512,提示数量 4)。
更好的研究,从现在开始
从论文设计到论文写作,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。