Sequential recommenders weight historical interactions either through positional self-attention as in Transformers or through a single implicit decay schedule as in State-Space Models. Neither makes the multi-scale temporal structure of real user behaviour explicit. We propose MARS, an encoder-agnostic aggregation operator that consumes real timestamps and produces K summaries emphasising distinct recency scales, fused by a context-adaptive gate. MARS adds at most 6% parameters and runs in $\mathcal{O}(LdK)$ time. MARS adapts to data density by automatically selecting between two encoder instantiations: MARS-T (Transformer) for sparse data and MARS-M (Mamba) for dense data, based on the average sequence length of the training set. On five public benchmarks against ten Transformer- and Mamba-based baselines under a unified RecBole protocol, MARS attains the best HR@10 on every benchmark, with mean relative gain +19.7% over the strongest content-only Transformer baseline on sparse data (reaching +36.2% on Games) and +3.2% HR@10 / +0.9% NDCG over SIGMA on dense ML-1M at 42% fewer MFLOPs, occupying the accuracy-efficiency Pareto frontier across the data-density spectrum. A backbone-only ablation isolates the marginal contribution of MARS at +4% to +19% HR@10 on sparse data and motivates the dual-instantiation design. The code is included in the supplementary material.

翻译：序列推荐模型通过Transformer中的位置自注意力或状态空间模型中的单一隐式衰减调度来权衡历史交互，但两者均未显式建模用户真实行为的多尺度时间结构。我们提出MARS——一种编码器无关的聚合算子，它利用真实时间戳生成K个强调不同新近尺度的摘要，并通过上下文自适应门控融合。MARS仅增加至多6%参数，运行时间复杂度为$\mathcal{O}(LdK)$。该算子可根据训练集平均序列长度自动适应数据密度，在两种编码器实例化间切换：面向稀疏数据的MARS-T（基于Transformer）与面向稠密数据的MARS-M（基于Mamba）。在统一RecBole协议下，针对五个公开基准数据集与十个Transformer及Mamba基线模型，MARS在所有基准上取得最佳HR@10，相较于最强内容无关Transformer基线在稀疏数据上平均提升+19.7%（在Games数据集达+36.2%），在稠密ML-1M上以42%更少MFLOPs实现HR@10/NDCG分别超越SIGMA +3.2%/+0.9%，占据数据密度谱系下的精度-效率帕累托前沿。仅含主干网络的消融实验表明，MARS在稀疏数据上贡献额外+4%至+19%的HR@10，并验证了双实例化设计的合理性。代码包含在补充材料中。