01.1 - Graduação (Sede)
URI permanente desta comunidadehttps://arandu.ufrpe.br/handle/123456789/2
Navegar
2 resultados
Resultados da Pesquisa
Item Aprendizado profundo com capacidade computacional reduzida: uma aplicação à quebra de CAPTCHAs(2018-08-16) Melo, Diogo Felipe Félix de; Sampaio, Pablo Azevedo; http://lattes.cnpq.br/8865836949700771; http://lattes.cnpq.br/2213650736070295During the last decade, Deep Neural Networks has been shown to be a powerfull machine learn technique. Generally, to obtain relevant results, these techniques require high computacional power and large volumes of data, which can be a limiting factor on some cases. Neverthless, a careful project of trainig and archtecture may help to reduce these requirements. In the this work we present a comparative approach to the application of deep neural networks to text based CAPTCHAs as a way to cope with these limitations. We studied models that are capable of learn to segment and identify the text content of images, only based on examples. By experimentation of different hiper-parameters and architectures, we were capable to obtain a final model with 96.06% of token prediction accuracy in approximately 3 hours of training in a simple personal computer.Item Extração de assinaturas de identificação de memória flash baseado em PUF com Arduino e validação com Perceptron Multicamadas(2019) Viana, Caio Bezerra; Araújo, Carlos Julian Menezes; http://lattes.cnpq.br/3156174527107999; http://lattes.cnpq.br/8717197088764951Device’s identity management is considered a core component of IoT security. Themain authentication methods use the concept of cryptographic key, this means that these curity provided by encryption is directly related to the key’s secrecy capacity. If thiskey is known by an intruder, the entire communication process is compromised sincethe messages’ content can be decrypted. The manufacture of some electronic devices may influence their physical behavior due to the existence of uncontrollable variables in-herent in the manufacturing process. Physical Unclonable Functions (PUF) techniquescan use these variables as a source for generating chip’s identification signatures. Thiswork proposes a flash memory identification approach, due to the wides pread use ofthis type of memory in current mobile devices, which use a memory block sequence verification technique based on the Program Operation Latency technique. With thisaim, the Arduino platform was used as a tool for extracting these identification signa-tures, togheter with a validation of the signatures conducted by a MultiLayer Perceptron(MLP). Who was able to learn enough about these signatures and generalize in the fu-ture, and correctly classified the classes of identification signatures used in the tests,thus supporting a flash memory’s differentiation mechanism.