Adversarial training (AT) is currently one of the most successful methods to obtain adversarial robustness deep neural networks. However, phenomenon robust overfitting, i.e., starts decrease significantly during AT, has been problematic, not only making practitioners consider a bag tricks for training, e.g., early stopping, but also incurring significant generalization gap in robustness. In thi...

Abstract The exponentially growing number of scientific papers stimulates a discussion on the interplay between quantity and quality in science. In particular, one may wonder which publication strategy offer more chances success: publishing lots papers, producing few hit or something between. Here we tackle this question by studying portfolios Nobel Prize laureates. A comparative analysis diffe...

We study time-consistency of optimization problems, where we say that an problem is time-consistent if its optimal solution, or the policy for choosing actions, does not depend on when solved. Time-consistency a minimal requirement decisions made based solution to be rational. show return can gain by taking "optimal" actions selected solving time-inconsistent surely dominated could "suboptimal"...

Recent work has increased the performance of Generative Adversarial Networks (GANs) by enforcing a consistency cost on discriminator. We improve this technique in several ways. first show that regularization can introduce artifacts into GAN samples and explain how to fix issue. then propose modifications procedure designed its performance. carry out extensive experiments quantifying benefit our...

The concept of community detection has long been used as a key device for handling the mesoscale structures in networks. Suitably conducted reveals various embedded informative substructures network topology. However, regarding practical usage detection, it always tricky problem to assign reasonable resolution networks interest. Because absence unanimously accepted criterion, most previous stud...