Learning-by-examples techniques as applied to electromagnetics

Abstract

There is a wide number of problems in electromagnetic (EM) engineering that require a real-time response or in which the input–output relationship is not a-priori known or cannot be defined due to the complexity of the scenario at hand. These issues have induced researchers toward the development and application of Learning-by-Examples (LBE) techniques thanks to their extremely high computational efficiency and to their capability to emulate the behavior of complex systems on the basis of a set of collected examples. In this framework, this paper aims to present an overview of the state-of-the-art and recently developed LBE-based strategies as applied to the solution of engineering problems in the field of Electromagnetics. Starting from a general and basic introduction to LBE techniques, the most popular LBE methods used in EM engineering will be re-called. Afterwards, a review on how the LBE methodologies have been applied by various researchers in different application contexts, including among others antennas, microwave circuits, inverse scattering and remote sensing, will be given. Finally, current research trends and envisaged developments are discussed.

Keywords:

• Learning-by-examples
• artificial neural networks
• support vector machines
• Gaussian processes
• engineering electromagnetics
• antennas
• direction-of-arrival estimation
• microwave circuits
• electromagnetic scattering
• remote sensing
• electromagnetic compatibility

Notes

No potential conflict of interest was reported by the authors.

1 The curse of dimensionality is the problem related to the fact that the size of the training set needs to exponentially grow with the dimensions of training samples to maintain a constant sampling density of the input space. Differently, the over-fitting arises when the surrogate function matches the training data with a too high degree of accuracy.

2 The extension to multi-class classification problems is straightforward and in that case is a discrete (e.g. integer) value [8].

Additional information

Funding

This work benefited from the networking activities carried out within the SNATCH Project (2017–2019) funded by the Italian Ministry of Foreign Affairs and International Cooperation, Directorate General for Cultural and Economic Promotion and Innovation and within the Project Zero Energy Buildings in Smart Urban Districts (2014–2017) funded by the Italian Ministry of Education, University, and Research under Grant CTN01_00034_594053 of the National Technological Cluster on Smart Communities.