Guiding the retraining of convolutional neural networks against adversarial inputs.

Background: When using deep learning models, one of the most critical vulnerabilities is their exposure to adversarial inputs, which can cause wrong decisions (., incorrect classification of an image) with minor perturbations. To address this vulnerability, it becomes necessary to retrain the affected model against adversarial inputs as part of the software testing process.

Towards suitable description of reference architectures.

Due to the increasing size and complexity of many current software systems, the architectural design of these systems has become a considerately complicated task. In this scenario, reference architectures have already proven to be very relevant to support the architectural design of systems in diverse critical application domains, such as health, avionics, transportation, and the automotive sector. However, these architectures are described in many different approaches, such as using textual description, informal models, and even modeling languages as UML.