← 研踪 PaperTrailarXiv:2610.23135 · v2
Wait to Learn: Timing the Gradient Return in Over-the-Air Split Learning over an Aging Single-Antenna Channel
首次提交 2026-10-16 16:35
版本更新 2026-10-26 16:34
雷达观测 2026-10-29 00:36
摘要
In over-the-air split learning over an aging single-antenna Rayleigh channel, the mean returned gradient equals the antenna-domain gradient scaled by the temporal correlation. Uncorrected training collapses in negative-correlation lobes. Near a correlation zero, no nonzero scalar calibration improves the ratio of mean power to innovation power for an individual transmission, and, on the tested grid, using sixteen times as many pilots recovers only 7 percent of the training gap. Returning at the first correlation extremum with phase correction, 1.3 ms later at 3.5 GHz and 15 m/s, achieves 90.2 percent accuracy, compared with 91.8 percent with CSI. In the tested line-of-sight setting, the near-zero gap becomes negligible.
作者备注：revised: theory reframed as instantiation of classical results, added line-of-sight extension
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-10-26 16:34 · 最新观测版本
Wait to Learn: Timing the Gradient Return in Over-the-Air Split Learning over an Aging Single-Antenna Channel
In over-the-air split learning over an aging single-antenna Rayleigh channel, the mean returned gradient equals the antenna-domain gradient scaled by the temporal correlation. Uncorrected training collapses in negative-correlation lobes. Near a correlation zero, no nonzero scalar calibration improves the ratio of mean power to innovation power for an individual transmission, and, on the tested grid, using sixteen times as many pilots recovers only 7 percent of the training gap. Returning at the first correlation extremum with phase correction, 1.3 ms later at 3.5 GHz and 15 m/s, achieves 90.2 percent accuracy, compared with 91.8 percent with CSI. In the tested line-of-sight setting, the near-zero gap becomes negligible.
查看此版本解读 →查看此版本原文 ↗