RFD：用共享算子分块生成长时程预测，参数仅0.08–0.46M
Compose the Horizon: Parameter-Shared Blockwise Forecasting under a Sequential-Compute Trade-off
arXiv上新出的时序预测论文，用参数共享算子分块滚动预测，0.46M参数就赢了24组里的23组，不过上下文比基线长，要仔细看对比设置。
论文提出 Recursive Functional Decomposition（RFD），将长时程预测分解为三个参数共享的块预测，每块预测结果滚动加入上下文再预测下一块。实验中移除前向反馈在全部四个数据集上都降低精度，而阻断其梯度几乎无影响，说明关键信号是不断更新的预测状态本身。RFD 参数量仅 0.08–0.46M，在 24 组对照中 23 组取得更低测试误差，但其对比采用跨协议设置：RFD 用 512/720 步上下文，十个基线家族仅用 96 步。高通道数下其串行执行比迭代基线更慢且更耗内存。
Compose the Horizon: Parameter-Shared Blockwise Forecasting under a Sequential-Compute Trade-off
Most long-horizon forecasters emit the complete future in one pass. We ask whether a compact operator can instead be reused to emit contiguous future blocks, trading stored parameters for sequential computation. Recursive Functional Decomposition (RFD) factorizes the horizon mapping into three parameter-shared block predictions. The blocks occupy disjoint future intervals and are concatenated, not added as trend, seasonal, or residual components, and each emitted block joins the rolling context that conditions the next application. In controlled three-seed experiments, removing forward feedback hurts all four evaluated datasets while stopping gradients through it changes little, so the useful signal is the evolving forecast state itself; reapplying the shared operator helps consistently, but by less than widening it. The model uses 0.08--0.46M parameters, yet its sequential channel-wise execution is slower and more memory-intensive than an iterative baseline at high channel count. In a frozen long-context evaluation against ten strong baseline families, RFD records lower test error in 23 of 24 cells; the comparison is cross-protocol, with RFD using 512/720-step context and the baselines using 96.
