The Experts below are selected from a list of 39972 Experts worldwide ranked by ideXlab platform

A. Sierra-alonso - One of the best experts on this subject based on the ideXlab platform.

  • WICSA - Heuristics for the transition from analysis to software architecture
    Proceedings. Fourth Working IEEE IFIP Conference on Software Architecture (WICSA 2004), 1
    Co-Authors: J.e. Perez-martinez, A. Sierra-alonso
    Abstract:

    To our knowledge, no current software development methodology explicitly describes how to transit between the different development stages it proposes. This is more evident in the transition from the analysis stage to the software architecture of the application. This paper presents the first semiautomatic method to derive the software architecture of a system from its analysis. The proposal is a set of heuristics that, when applied to the analysis artifacts, generate the software architecture of the application. This proposal has several benefits: (1) the software architecture of the system is directly derived by applying the heuristics; (2) there is a direct Trace Relationship between the analysis artifacts and the elements of the resulting architecture, which eases the system maintenance.

J.e. Perez-martinez - One of the best experts on this subject based on the ideXlab platform.

  • WICSA - Heuristics for the transition from analysis to software architecture
    Proceedings. Fourth Working IEEE IFIP Conference on Software Architecture (WICSA 2004), 1
    Co-Authors: J.e. Perez-martinez, A. Sierra-alonso
    Abstract:

    To our knowledge, no current software development methodology explicitly describes how to transit between the different development stages it proposes. This is more evident in the transition from the analysis stage to the software architecture of the application. This paper presents the first semiautomatic method to derive the software architecture of a system from its analysis. The proposal is a set of heuristics that, when applied to the analysis artifacts, generate the software architecture of the application. This proposal has several benefits: (1) the software architecture of the system is directly derived by applying the heuristics; (2) there is a direct Trace Relationship between the analysis artifacts and the elements of the resulting architecture, which eases the system maintenance.

Zidong Wang - One of the best experts on this subject based on the ideXlab platform.

  • A Trace-restricted Kronecker-Factored Approximation to Natural Gradient.
    arXiv: Learning, 2020
    Co-Authors: Kai-xin Gao, Xiao-lei Liu, Zheng-hai Huang, Min Wang, Zidong Wang
    Abstract:

    Second-order optimization methods have the ability to accelerate convergence by modifying the gradient through the curvature matrix. There have been many attempts to use second-order optimization methods for training deep neural networks. Inspired by diagonal approximations and factored approximations such as Kronecker-Factored Approximate Curvature (KFAC), we propose a new approximation to the Fisher information matrix (FIM) called Trace-restricted Kronecker-factored Approximate Curvature (TKFAC) in this work, which can hold the certain Trace Relationship between the exact and the approximate FIM. In TKFAC, we decompose each block of the approximate FIM as a Kronecker product of two smaller matrices and scaled by a coefficient related to Trace. We theoretically analyze TKFAC's approximation error and give an upper bound of it. We also propose a new damping technique for TKFAC on convolutional neural networks to maintain the superiority of second-order optimization methods during training. Experiments show that our method has better performance compared with several state-of-the-art algorithms on some deep network architectures.

Kai-xin Gao - One of the best experts on this subject based on the ideXlab platform.

  • A Trace-restricted Kronecker-Factored Approximation to Natural Gradient.
    arXiv: Learning, 2020
    Co-Authors: Kai-xin Gao, Xiao-lei Liu, Zheng-hai Huang, Min Wang, Zidong Wang
    Abstract:

    Second-order optimization methods have the ability to accelerate convergence by modifying the gradient through the curvature matrix. There have been many attempts to use second-order optimization methods for training deep neural networks. Inspired by diagonal approximations and factored approximations such as Kronecker-Factored Approximate Curvature (KFAC), we propose a new approximation to the Fisher information matrix (FIM) called Trace-restricted Kronecker-factored Approximate Curvature (TKFAC) in this work, which can hold the certain Trace Relationship between the exact and the approximate FIM. In TKFAC, we decompose each block of the approximate FIM as a Kronecker product of two smaller matrices and scaled by a coefficient related to Trace. We theoretically analyze TKFAC's approximation error and give an upper bound of it. We also propose a new damping technique for TKFAC on convolutional neural networks to maintain the superiority of second-order optimization methods during training. Experiments show that our method has better performance compared with several state-of-the-art algorithms on some deep network architectures.

Xiao-lei Liu - One of the best experts on this subject based on the ideXlab platform.

  • A Trace-restricted Kronecker-Factored Approximation to Natural Gradient.
    arXiv: Learning, 2020
    Co-Authors: Kai-xin Gao, Xiao-lei Liu, Zheng-hai Huang, Min Wang, Zidong Wang
    Abstract:

    Second-order optimization methods have the ability to accelerate convergence by modifying the gradient through the curvature matrix. There have been many attempts to use second-order optimization methods for training deep neural networks. Inspired by diagonal approximations and factored approximations such as Kronecker-Factored Approximate Curvature (KFAC), we propose a new approximation to the Fisher information matrix (FIM) called Trace-restricted Kronecker-factored Approximate Curvature (TKFAC) in this work, which can hold the certain Trace Relationship between the exact and the approximate FIM. In TKFAC, we decompose each block of the approximate FIM as a Kronecker product of two smaller matrices and scaled by a coefficient related to Trace. We theoretically analyze TKFAC's approximation error and give an upper bound of it. We also propose a new damping technique for TKFAC on convolutional neural networks to maintain the superiority of second-order optimization methods during training. Experiments show that our method has better performance compared with several state-of-the-art algorithms on some deep network architectures.