Computing Based on Material Training: Application to Binary Classification Problems

Abstract

any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Additional information:Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. Abstract-Evolution-in-materio is a form of unconventional computing combining materials' training and evolutionary search algorithms. In previous work, a mixture of single-walled-carbonnanotubes (SWCNTs) dispersed in a liquid crystal (LC) was trained so that its morphology and electrical properties were gradually changed to perform a computational task. Materialbased computation is treated as an optimisation problem with a hybrid search space consisting of the voltages used for creating the electrical field and the material's infinitely possible SWCNT arrangements in LC. In this paper, we study solutions using synthetic data with a non-linear separating boundary. In addition, results for two real life datasets with partly merged classes are presented. The training process is based on a differential evolution (DE) algorithm, which subjects the SWCNT/LC material to repeated electrical charging, leading to progressive morphological and electric conductivity modifications. It is shown that the material configuration the DE algorithm converges to form a non-negligible part of the solution. Furthermore, the problem's complexity is relevant to the properties of the resulting physical solver. The material structures created when training for a problem allow the retraining for a less complex one. The result is a doubly-trained material that keeps the memory of the original more complex problem. This is not the case for doubly-trained materials where initial training is for the less complex problem. The optimal electric field found by the DE algorithm is also a necessary solution component for the material's output to be interpreted as a computation.

Confidence Measures for Carbon-Nanotube / Liquid Crystals Classifiers

“…The contribution of the SWCNT structures transformed by the training process have been assessed by sending unseen instances to the evolved device, with no configuration voltages [17]. Finally, the stability of these structures when subjected to retraining was discussed in [18]. The first three columns of Table II present the optimal training Φ * e and mean verification Φ v e errors averaged over 10 experiments for the three datasets, as well as the standard deviation σ of results across experiments.…”

“…It has been observed in [16] that SWCNTs dispersed in LCs tend to aggregate along an applied electric field, forming percolation paths between electrodes. Varying this electric field results in modification of these percolation paths and the formation of complex SWCNT structures in the composite which favour computation [17,18]. At the start of each experiment, a sample of the SWCNT/LC composite was drop-cast within a nylon washer of 2.5 mm internal diameter.…”

Evaluation of Contributing Factors Affecting Number of Vehicles Involved in Crashes Using Machine Learning Techniques in Rural Roads of Cosenza, Italy

“…Finally, using the least-squares process from multiple regression analysis, a normal equation is achieved based on Eq 8, which calculates the vector of the best coefficients for Eq 4 [78].…”

Towards Intelligently Designed Evolvable Processors

“…As previously mentioned, different nanomaterials can be selected for a wide range of different properties, potentially allowing for fast, power efficient unconventional computing. While some nanomaterials can alter their internal connections (e.g., memristive materials (Strukov et al, 2008), and nanomaterials suspended in solution (Vissol-Gaudin et al, 2017)), these EiM processors lie outside the scope of this article. Instead, we focus on 'fixed' material network EiM processors, enabling their foundational operating principles to be established.…”