This repository functions as an archive of all code written for my Master's dissertation. A PDF of the entire dissertation book is included in the root of this code repository.
"We present a framework to evaluate the computational capabilities of plant physiological processes in plant models for physical reservoir computing (PRC). PRC is a computing paradigm that enables the use of physical substrates for computation. Recently, the use of live plants in a PRC context was demonstrated. We expand upon this work in silico using plant models for grapevine and wheat.
In our experiments, we investigated the use of leaf temperature, transpiration rate, photosynthesis rate, light absorption, water flow, and hydraulic pressure as readout traits. Each of these traits displayed reservoir dynamics and correlated well with biologically relevant tasks. Surprisingly, our framework highlighted unusual behavior in mechanistic plant models, which the authors did not yet report.
We propose that the plant-modeling community can adapt the methods in this work to verify the behavior of plant models. We believe that this is just the starting point for plant reservoir computing. If knowledge and technology can scale, there will be countless opportunities in agricultural technology."
(Excerpt from the extended abstract)