BZPEERSource author record
Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.
Catalog footprint
Published work
We establish standard zero-free regions with no exceptional Landau-Siegel zeros for Rankin-Selberg L-functions and triple product L-functions in several new families for which modularity is not yet known.
Let F be a number field. Let π1,π2 be cuspidal automorphic representations of GL2(AF), and let π be a cuspidal automorphic representation of either GL2(AF) or GL3(AF). When (π1,π2,π) is of general type, we show that the triple product L-function L(s,π1×π2×π) on either GL(2)×GL(2)×GL(2) or GL(2)×GL(2)×GL(3) has a standard zero-free region with no exceptional Landau-Siegel zero. Moreover, when (π1,π2,π) is not of general type, we give precise conditions when L(s,π1×π2×π) could possibly have exceptional Landau-Siegel zeros.
Many modern machine learning algorithms such as generative adversarial networks (GANs) and adversarial training can be formulated as minimax optimization. Gradient descent ascent (GDA) is the most commonly used algorithm due to its simplicity. However, GDA can converge to non-optimal minimax points. We propose a new minimax optimization framework, GDA-AM, that views the GDAdynamics as a fixed-point iteration and solves it using Anderson Mixing to con-verge to the local minimax. It addresses the diverging issue of simultaneous GDAand accelerates the convergence of alternating GDA. We show theoretically that the algorithm can achieve global convergence for bilinear problems under mild conditions. We also empirically show that GDA-AMsolves a variety of minimax problems and improves GAN training on several datasets
Research graph
Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.
BZPEER is loading the nearby papers, people, topics and institutions for this page.