Evolution Of Activation Functions for Neural Architecture Search by Nader, Andrew

“…However, to the best of our knowledge, the design of new activation functions has mostly been done by hand. In this work, we propose the use of a self-adaptive evolutionary algorithm that searches for new activation functions using a genetic programming approach, and we compare the performance of the obtained activation functions to ReLU. …”
Get full text
Get full text
Get full text

R-CONV++: uncovering privacy vulnerabilities through analytical gradient inversion attacks by Tamer Ahmed Eltaras (22565414)

“…The first algorithm presents a novel data leakage method that efficiently exploits convolutional layer gradients, demonstrating that even with non-fully invertible activation functions, such as ReLU, training samples can be analytically reconstructed directly from gradients without the need to reconstruct intermediate layer outputs. …”