Zobrazeno 1 - 2
of 2
pro vyhledávání: '"Raval, Param"'
Adversarial example (AE) aims at fooling a Convolution Neural Network by introducing small perturbations in the input image.The proposed work uses the magnitude and phase of the Fourier Spectrum and the entropy of the image to defend against AE. We d
Externí odkaz:
http://arxiv.org/abs/2110.13935
This study aims to provide a comparative analysis of performance of certain models popular in machine learning and the BERT model on the Stanford Question Answering Dataset (SQuAD). The analysis shows that the BERT model, which was once state-of-the-
Externí odkaz:
http://arxiv.org/abs/2005.11313