Факультет електроніки та інформаційних технологій (ЕлІТ)
Permanent URI for this communityhttps://devessuir.sumdu.edu.ua/handle/123456789/20
Browse
2 results
Search Results
Item Model and Training Method of the Resilient Image Classifier Considering Faults, Concept Drift, and Adversarial Attacks(MDPI, 2022) Москаленко, В`ячеслав Васильович; Москаленко, Вячеслав Васильевич; Moskalenko, Viacheslav Vasylovych; Kharchenko, V.; Москаленко, Альона Сергіївна; Москаленко, Алена Сергеевна; Moskalenko, Alona Serhiivna; Петров, Сергій Олександрович; Петров, Сергей Александрович; Petrov, Serhii OleksandrovychModern trainable image recognition models are vulnerable to different types of perturbations; hence, the development of resilient intelligent algorithms for safety-critical applications remains a relevant concern to reduce the impact of perturbation on model performance. This paper proposes a model and training method for a resilient image classifier capable of efficiently functioning despite various faults, adversarial attacks, and concept drifts. The proposed model has a multi-section structure with a hierarchy of optimized class prototypes and hyperspherical class boundaries, which provides adaptive computation, perturbation absorption, and graceful degradation. The proposed training method entails the application of a complex loss function assembled from its constituent parts in a particular way depending on the result of perturbation detection and the presence of new labeled and unlabeled data. The training method implements principles of self-knowledge distillation, the compactness maximization of class distribution and the interclass gap, the compression of feature representations, and consistency regularization. Consistency regularization makes it possible to utilize both labeled and unlabeled data to obtain a robust model and implement continuous adaptation. Experiments are performed on the publicly available CIFAR-10 and CIFAR-100 datasets using model backbones based on modules ResBlocks from the ResNet50 architecture and Swin transformer blocks. It is experimentally proven that the proposed prototype-based classifier head is characterized by a higher level of robustness and adaptability in comparison with the dense layer-based classifier head. It is also shown that multi-section structure and self-knowledge distillation feature conserve resources when processing simple samples under normal conditions and increase computational costs to improve the reliability of decisions when exposed to perturbations.Item Comparison of Morphology and Cytotoxicity of Cellulose Composites with Nano- and Microhydroxyapatite for Bone Tissue Engineering(Sumy State University, 2015) Liesiene, J.; Baniukaitiene, O.; Babenko, N.; Harkavenko, V.; Kharchenko, V.Scaffolds for bone tissue regeneration must be precisely designed as they support cell attachment, proliferation, differentiation and blood vessel in-growth. Moreover, the materials which are implanted in human body must be non-cytotoxic, absolutely harmless. In this work cellulose scaffolds with nanohydroxyapatite and microhydroxyapatite were prepared. The results obtained in this work revealed that the morphology of the scaffolds depended on the size of hydroxyapatite particles. The porosity of the scaffolds varied from 66 to 72%. The pores were interconnected with the average diameter 0.49 and 0.54 mm for scaffolds with nano- and microhydroxyapatite, respectively. Biocompatibility and potential toxicity of the experimental cellulose/hydroxyapatite scaffolds were tested. It was determined that the scaffolds containing nanohydroxyapatite particles showed slight cytotoxic effect.