# In-situ Physical Adjoint Computing in multiple-scattering electromagnetic environments for wave control

arXiv: https://arxiv.org/abs/2503.21107 · HTML: https://mlsanity.com/paper/2503.21107/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Mar 2025.

## Built on by (most cited first, 1 of 1)

- [Physical policy gradient theorem for in situ stochastic-adjoint training](https://mlsanity.com/paper/2609.05808/) (2026)

## Abstract (arXiv)

Controlling electromagnetic wave propagation in multiple scattering systems is a challenging endeavor due to the extraordinary sensitivity generated by strong multi-path contributions at any given location. Overcoming such complexity has emerged as a central research theme in recent years, motivated both by a wide range of applications -- from wireless communications and imaging to optical micromanipulations -- and by the fundamental principles underlying these efforts. Here, we show that an it in-situ manipulation of the myriad scattering events, achieved through time- and energy-efficient adjoint optimization (AO) methodologies, enables it real time wave-driven functionalities such as targeted channel emission, coherent perfect absorption, and camouflage. Our paradigm shift exploits the highly multi-path nature of these complex environments, where repeated wave-scattering dramatically amplifies small local AO-informed system variations. Our approach can be immediately applied to in-door wireless technologies and incorporated into diverse wave-based frameworks including imaging, power electronic and optical neural networks.