منابع مشابه
Diagonal Rescaling For Neural Networks
We define a second-order neural network stochastic gradient training algorithm whose block-diagonal structure effectively amounts to normalizing the unit activations. Investigating why this algorithm lacks in robustness then reveals two interesting insights. The first insight suggests a new way to scale the stepsizes, clarifying popular algorithms such as RMSProp as well as old neural network t...
متن کاملRescaling the Energy Function in Hopfield Networks
In this paper we propose an approach that rescales the distance matrix of the energy function in the Hoppeld network for solving optimization problems. We rescale the distance matrix by normalizing each row in the matrix and then adjusting the parameter for the distance term. This scheme has the capability of reducing the eeects of clustering in data distributions, which is one of main reasons ...
متن کاملFast Training of Convolutional Neural Networks via Kernel Rescaling
Training deep Convolutional Neural Networks (CNN) is a time consuming task that may take weeks to complete. In this article we propose a novel, theoretically founded method for reducing CNN training time without incurring any loss in accuracy. The basic idea is to begin training with a pre-train network using lower-resolution kernels and input images, and then refine the results at the full res...
متن کاملLocal structures determine performance within complex networks
Networks are ubiquitous. We as individuals are part of various social networks and each of us depends on multiple communication, traffic and supply networks in our everyday life. We are, however, still far from completely understanding and controlling those networks. Network motifs, few nodes (un)directed subgraphs, are a well-defined intermediate scale for characterizing the local structure of...
متن کاملTowards Understanding of Complex Communication Networks: Performance, Phase Transitions & Control
The paper discusses a possibility of phase transitions and metastability in various types of complex communication networks as well as implication of these phenomena for network performance evaluation and control. Specific cases include connection-oriented networks with dynamic routing, TCP/IP networks under random flow arrivals/departures, and multiservice wireless cellular networks. Despite r...
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: IEEE Communications Letters
سال: 2014
ISSN: 1089-7798
DOI: 10.1109/lcomm.2014.021014.132684