Reciprocity Separates Gradient Flow from Rotation in Conservative Physical Learning
Abstract
Physical learning lets a trainable material or network use its own physical response to carry error signals, reducing the need for a separately programmed backward computation. We ask what determines whether such a system follows conventional gradient descent or evolves along a genuinely different learning trajectory. Our canonical model is a directed layered transport network in which every node redistributes a fixed amount of flow, so learning preserves positivity and total mass. In this model, conservation constrains only the allowable learning directions. Within the matched response class studied here, adjoint matching gives the physical output response a symmetric form. Non-negative mode-wise feedback then produces a reciprocal closed-loop response and a reweighted gradient flow. Adding an antisymmetric boundary component makes the closed-loop response rotational: the learning path can turn while the error driving that update still decreases at that moment. Turning is not automatically beneficial. Its finite-step effect is set by local curvature, and its accumulated effect also depends on step selection and on the new states visited along the path. Numerical consistency checks reproduce the exact response structure, predict the sign of the local effect across new network families, and show how trajectory drift can negate a local advantage. These results separate the roles of conservation, reciprocity, and nonreciprocity in physical learning.
Related
- Symmetry-driven in-situ physical adjoint computing beyond reciprocity
- Conservative Numerical Construction of Vlasov-Maxwell Equilibria and an Exact Cancellation of E× B Flow by an Electrostatic Pressure-Gradient Flow
- Physical policy gradient theorem for in situ stochastic-adjoint training
- Spin-Hall devices: spin relaxation spatially separates current injection from Joule dissipation
- Scrambling Separates Superradiance from Quantum Chaos
BibTeX
@misc{niu2026reciprocity,
title = {{Reciprocity Separates Gradient Flow from Rotation in Conservative Physical Learning}},
author = {Ruiwu Niu and Xiaowen Bi and Michaël Antonie van Wyk},
howpublished = {arXiv},
year = {2026},
month = aug,
doi = {10.48550/arxiv.2608.30778},
eprint = {2608.30778},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.30778},
}
RIS
TY - UNPB
TI - Reciprocity Separates Gradient Flow from Rotation in Conservative Physical Learning
AU - Ruiwu Niu
AU - Xiaowen Bi
AU - Michaël Antonie van Wyk
PY - 2026
DA - 2026/08/31
PB - arXiv
DO - 10.48550/arxiv.2608.30778
UR - https://arxiv.org/abs/2608.30778
AB - Physical learning lets a trainable material or network use its own physical response to carry error signals, reducing the need for a separately programmed backward computation. We ask what determines whether such a system follows conventional gradient descent or evolves along a genuinely different learning trajectory. Our canonical model is a directed layered transport network in which every node redistributes a fixed amount of flow, so learning preserves positivity and total mass. In this model, conservation constrains only the allowable learning directions. Within the matched response class studied here, adjoint matching gives the physical output response a symmetric form. Non-negative mode-wise feedback then produces a reciprocal closed-loop response and a reweighted gradient flow. Adding an antisymmetric boundary component makes the closed-loop response rotational: the learning path can turn while the error driving that update still decreases at that moment. Turning is not automatically beneficial. Its finite-step effect is set by local curvature, and its accumulated effect also depends on step selection and on the new states visited along the path. Numerical consistency checks reproduce the exact response structure, predict the sign of the local effect across new network families, and show how trajectory drift can negate a local advantage. These results separate the roles of conservation, reciprocity, and nonreciprocity in physical learning.
ER -