Work

Improving robustness of Transformers for power quality disturbance classification via optimized relevance maps

Year: 2025

Type: article

Abstract: Power Quality Disturbances (PQDs) are a critical area of research in the power systems domain. The growing complexity of modern power systems increase the the variability of PQDs that may occur, making it challenging to manage power quality using traditional methods, and motivate inclusion of machine learning models. In this paper, we propose a new approach to increase the robustness of a family of PQD classifiers against adversarial attacks, by increasing the explainability of the contextual relations between different features of the input signal, without adversarial training. To this end, we use the Transformer encoder model for PQD classification,relying on its self-attention mechanism to introduce a relevance property to this model. We show that by incorporating the resulting explainable features in the form of relevance maps into the loss function during the fine-tuning stage, we achieve improvement in the robustness and resiliency of the model against adversarial attacks, without compromising accuracy, and without investing additional computational resources required for adversarial training. We also show that the proposed Transformer encoder method is more adversarial robust in comparison to state-of-the-art Convolutional Neural Network (CNN).

Source: Engineering Applications of Artificial Intelligence

DOI: 10.1016/j.engappai.2025.112138

Authors: Itamar Kapuza, Elinor Ginzburg-Ganz, Ram Machlev, Yoash Levron

Institution: Technion – Israel Institute of Technology

Language: English

Topic: Power Quality and Harmonics

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: hybrid

Estimated APC paid: $3,200