Optimal Boltzmann hierarchies with nonvanishing spatial curvature
نویسندگان
چکیده
منابع مشابه
Learning Feature Hierarchies with Centered Deep Boltzmann Machines
Deep Boltzmann machines are in principle powerful models for extracting the hierarchical structure of data. Unfortunately, attempts to train layers jointly (without greedy layerwise pretraining) have been largely unsuccessful. We propose a modification of the learning algorithm that initially recenters the output of the activation functions to zero. This modification leads to a better condition...
متن کاملDeep Boltzmann Machines as Feed-Forward Hierarchies
The deep Boltzmann machine is a powerful model that extracts the hierarchical structure of observed data. While inference is typically slow due to its undirected nature, we argue that the emerging feature hierarchy is still explicit enough to be traversed in a feedforward fashion. The claim is corroborated by training a set of deep neural networks on real data and measuring the evolution of the...
متن کاملZero Curvature Formulations of Dual Hierarchies
Zero curvature formulations are given for the “dual hierarchies” of standard soliton equation hierarchies, recently introduced by Olver and Rosenau, including the physically interesting Fuchssteiner-Fokas-CamassaHolm hierarchy. PACS Numbers: 03.40.Kf, 47.20.Ky, 52.35.Sb Through the years since the discovery of the notion of “integrability” in PDEs, quite a number of integrable PDEs have been di...
متن کاملRobust Optimal Designs with Reduced Curvature
Researchers often find that nonlinear regression models are more relevant for their studies than are linear ones, and are thus often in a position of requiring efficient designs for a given nonlinear model. A common shortcoming of most optimal designs for nonlinear models used in practical settings, however, is that these designs typically focus only on first-order parameter variance or predict...
متن کاملGeometry of optimal path hierarchies
We investigate the hierarchy of optimal paths in a disordered landscape, based on the best path, the second best path and so on in terms of an energy. By plotting each path at a height according to its energy above some zero level, a landscape appears. This landscape is self-affine and controlled by two Hurst exponents: the one controlling the height fluctuations is 1/3 and the one controlling ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Physical Review D
سال: 2020
ISSN: 2470-0010,2470-0029
DOI: 10.1103/physrevd.102.023511