Sign in

Effects of Loss Functions And Target Representations on Adversarial Robustness

By Sean Saito and Sujoy Roy
Understanding and evaluating the robustness of neural networks under adversarial settings is a subject of growing interest. Attacks proposed in the literature usually work with models trained to minimize cross-entropy loss and output softmax probabilities. In this work, we present interesting experimental results that suggest the importance of considering other... Show more
February 8, 2019
Loading PDF…
Loading full text...
Similar articles
Loading recommendations...