The Experts below are selected from a list of 569205 Experts worldwide ranked by ideXlab platform

Jian Sun - One of the best experts on this subject based on the ideXlab platform.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    European Conference on Computer Vision, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the indirect metric of computation complexity, i.e., FLOPs. However, the direct metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical guidelines for efficient network Design. Accordingly, a new Architecture is presented, called ShuffleNet V2. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the \emph{indirect} metric of computation complexity, i.e., FLOPs. However, the \emph{direct} metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical \emph{guidelines} for efficient network Design. Accordingly, a new Architecture is presented, called \emph{ShuffleNet V2}. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.

Martina Keitsch - One of the best experts on this subject based on the ideXlab platform.

  • Sustainable Architecture, Design and Housing
    Sustainable Development, 2012
    Co-Authors: Martina Keitsch
    Abstract:

    This editorial article for the special issue of Sustainable Development ‘Sustainable Architecture, Design and housing’ introduces the main topics and reviews the following nine articles. The aim of the special issue is to discuss key concepts, methods, applications and lessons learned in sustainable Architecture, Design and housing. Sustainable Architecture challenges new and ingenious architectural Design at various levels. Examples are the following. (1) Minimizing the negative environmental impact of buildings by enhancing efficiency and moderating the use of materials, energy and development space. (2) Developing measures to relate form and adapt the Design to the site, the region and the climate. (3) Establishing a harmonious, long lasting relationship between the inhabitants and their surroundings by addressing the essence of good form-giving (Abidin et al., ). In summary, sustainable Architecture shall be well built, easy to use and beautiful. The selected articles highlight special aspects of sustainability in Architecture, Design and housing, either due to their novel theoretical contributions or in relation to tools, measurements and strategies. A main research objective on sustainable Design is to promote understanding for the particular problems that architects and Designers face and to discuss solutions to solve them. This can in turn also be described as the relevance of sustainable Design research for the overarching sustainable development debate. Sustainable Design is dynamic and comprises both normative, strategic, Designerly and technical aspects (Cross, ). Three areas are thematically highlighted in this special issue. * Overarching conception of sustainable Architecture and Design and their impact on a sustainable development. * Methodologies and tools and their applications to sustainable Design. * Policy and strategy development to assess and implement sustainable Design. Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment.

Xiangyu Zhang - One of the best experts on this subject based on the ideXlab platform.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    European Conference on Computer Vision, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the indirect metric of computation complexity, i.e., FLOPs. However, the direct metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical guidelines for efficient network Design. Accordingly, a new Architecture is presented, called ShuffleNet V2. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the \emph{indirect} metric of computation complexity, i.e., FLOPs. However, the \emph{direct} metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical \emph{guidelines} for efficient network Design. Accordingly, a new Architecture is presented, called \emph{ShuffleNet V2}. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.

Philippe Kruchten - One of the best experts on this subject based on the ideXlab platform.

  • DecidArch V2: An Improved Game to Teach Architecture Design Decision Making
    2019 IEEE International Conference on Software Architecture Companion (ICSA-C), 2019
    Co-Authors: Remco C. De Boer, Roberto Verdecchia, Patricia Lago, Philippe Kruchten
    Abstract:

    We report on the use of our DecidArch game to teach software Architecture Design decision making in two consecutive years. We compare the support of three learning goals for the first version of the game with the second, revised version. Results show how the game has clearly improved. For the remaining issues. we suggest final improvements.

  • decision making techniques for software Architecture Design a comparative survey
    ACM Computing Surveys, 2011
    Co-Authors: Davide Falessi, Giovanni Cantone, Rick Kazman, Philippe Kruchten
    Abstract:

    The Architecture of a software-intensive system can be defined as the set of relevant Design decisions that affect the qualities of the overall system functionality; therefore, architectural decisions are eventually crucial to the success of a software project. The software engineering literature describes several techniques to choose among architectural alternatives, but it gives no clear guidance on which technique is more suitable than another, and in which circumstances. As such, there is no systematic way for software engineers to choose among decision-making techniques for resolving tradeoffs in Architecture Design. In this article, we provide a comparison of existing decision-making techniques, aimed to guide architects in their selection. The results show that there is no “best” decision-making technique; however, some techniques are more susceptible to specific difficulties. Hence architects should choose a decision-making technique based on the difficulties that they wish to avoid. This article represents a first attempt to reason on meta-decision-making, that is, the issue of deciding how to decide.

  • a general model of software Architecture Design derived from five industrial approaches
    Journal of Systems and Software, 2007
    Co-Authors: Christine Hofmeister, Alexander Ran, Henk Obbink, Robert L Nord, Philippe Kruchten, Pierre America
    Abstract:

    We compare five industrial software Architecture Design methods and we extract from their commonalities a general software Architecture Design approach. Using this general approach, we compare across the five methods the artifacts and activities they use or recommend, and we pinpoint similarities and differences. Once we get beyond the great variance in terminology and description, we find that the five approaches have a lot in common and match more or less the ''ideal'' pattern we introduced. From the ideal pattern we derive an evaluation grid that can be used for further method comparisons.

Haitao Zheng - One of the best experts on this subject based on the ideXlab platform.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    European Conference on Computer Vision, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the indirect metric of computation complexity, i.e., FLOPs. However, the direct metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical guidelines for efficient network Design. Accordingly, a new Architecture is presented, called ShuffleNet V2. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.

  • shufflenet v2 practical guidelines for efficient cnn Architecture Design
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Xiangyu Zhang, Haitao Zheng, Jian Sun
    Abstract:

    Currently, the neural network Architecture Design is mostly guided by the \emph{indirect} metric of computation complexity, i.e., FLOPs. However, the \emph{direct} metric, e.g., speed, also depends on the other factors such as memory access cost and platform characterics. Thus, this work proposes to evaluate the direct metric on the target platform, beyond only considering FLOPs. Based on a series of controlled experiments, this work derives several practical \emph{guidelines} for efficient network Design. Accordingly, a new Architecture is presented, called \emph{ShuffleNet V2}. Comprehensive ablation experiments verify that our model is the state-of-the-art in terms of speed and accuracy tradeoff.