Search

Save for some special cases, current training methods for Generative Adversarial Networks (GANs) are at best guaranteed to converge to a ‘local Nash equilibrium’ (LNE). Such LNEs, however, can be arbitrarily far from an actual Nash equilibrium (NE), which implies that there are no guarantees on the quality of the found generator or classifier. This paper proposes to model GANs explicitly as finite games in mixed strategies, thereby ensuring that every LNE is an NE. We use the Parallel Nash Memory as a solution method, which is proven to monotonically converge to a resource-bounded Nash equilibrium. We empirically demonstrate that our method is less prone to typical GAN problems such as mode collapse and produces solutions that are less exploitable than those produced by GANs and MGANs., Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Interactive Intelligence

Save for some special cases, current training methods for Generative Adversarial Networks (GANs) are at best guaranteed to converge to a ‘local Nash equilibrium’ (LNE). Such LNEs, however, can be arbitrarily far from an actual Nash equilibrium (NE), which implies that there are no guarantees on the quality of the found generator or classifier. This paper proposes to model GANs explicitly as finite games in mixed strategies, thereby ensuring that every LNE is an NE. We use the Parallel Nash Memory as a solution method, which is proven to monotonically converge to a resource-bounded Nash equilibrium. We empirically demonstrate that our method is less prone to typical GAN problems such as mode collapse and produces solutions that are less exploitable than those produced by GANs and MGANs., Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Interactive Intelligence

Software is widely used, and society increasingly depends on its reliability. However, software has become so complex and it evolves so quickly that we fail to keep it under control. Therefore, we propose intents: fundamental laws that capture a software systems’ intended behavior (resilient, secure, safe, sustainable, etc.). The realization of this idea requires novel theories, algorithms, tools, and techniques to discover, express, verify, and evolve software intents. Thus, future software systems will be able to verify themselves that they meet their intents. Moreover, they will be able to respond to deviations from intents through selfcorrection. In this article we propose a research agenda, outlining which novel theories, algorithms and tools are required., Software Technology, Programming Languages

Software is widely used, and society increasingly depends on its reliability. However, software has become so complex and it evolves so quickly that we fail to keep it under control. Therefore, we propose intents: fundamental laws that capture a software systems’ intended behavior (resilient, secure, safe, sustainable, etc.). The realization of this idea requires novel theories, algorithms, tools, and techniques to discover, express, verify, and evolve software intents. Thus, future software systems will be able to verify themselves that they meet their intents. Moreover, they will be able to respond to deviations from intents through selfcorrection. In this article we propose a research agenda, outlining which novel theories, algorithms and tools are required., Software Technology, Programming Languages

Applied Sciences

Applied Sciences