← 研踪 PaperTrailarXiv:2604.20219 · v3
Layer-wise Geometric Approximation Rates for Deep Networks
首次提交 2026-04-22 14:09
版本更新 2026-11-16 23:54
雷达观测 2026-11-17 16:10
摘要
Depth is widely viewed as a central contributor to the success of deep neural networks, yet standard approximation results typically give error bounds only for the final output of a network. Inspired by classical multiscale expansions and multigrade deep learning, we give depth a scale-dependent interpretation by constructing a single network whose intermediate readouts approximate the target at successively finer scales. For any with and any integer , we construct a single network of width with the same activation pattern in every hidden layer, using sine in the first two neurons and in the rest. For every , the approximation error of the readout at layer is bounded by , where denotes the translation modulus of . Under the Hölder condition with and , this estimate yields the geometric bound . More generally, if has a known upper bound that tends to zero as , we obtain a stopping rule for adaptive depth selection. We can stop at the first depth where the error bound meets the desired tolerance and later append layers to meet a tighter tolerance while keeping all earlier maps and readouts fixed.
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。
v3 · 2026-11-16 23:54 · 最新观测版本
Layer-wise Geometric Approximation Rates for Deep Networks
Depth is widely viewed as a central contributor to the success of deep neural networks, yet standard approximation results typically give error bounds only for the final output of a network. Inspired by classical multiscale expansions and multigrade deep learning, we give depth a scale-dependent interpretation by constructing a single network whose intermediate readouts approximate the target at successively finer scales. For any with and any integer , we construct a single network of width with the same activation pattern in every hidden layer, using sine in the first two neurons and in the rest. For every , the approximation error of the readout at layer is bounded by , where denotes the translation modulus of . Under the Hölder condition with and , this estimate yields the geometric bound . More generally, if has a known upper bound that tends to zero as , we obtain a stopping rule for adaptive depth selection. We can stop at the first depth where the error bound meets the desired tolerance and later append layers to meet a tighter tolerance while keeping all earlier maps and readouts fixed.
查看此版本解读 →查看此版本原文 ↗