Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Mathematical Models to Measure the Variability of Nodes and Networks in Team Sports.

Entropy (Basel, Switzerland)·2021
Same author

Real-time 2D-3D door detection and state classification on a low-power device.

SN applied sciences·2021
Same author

Disruptive Mood Dysregulation Disorder: A New Diagnostic Approach to Chronic Irritability in Youth.

Focus (American Psychiatric Publishing)·2020
Same author

High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning.

Journal of biomolecular screening·2016
Same author

BIK-BUS: biologically motivated 3D keypoint based on bottom-up saliency.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2014
Same author

Disruptive mood dysregulation disorder: a new diagnostic approach to chronic irritability in youth.

The American journal of psychiatry·2014

相关实验视频

Updated: Jul 12, 2025

3D Scanning Technology Bridging Microcircuits and Macroscale Brain Images in 3D Novel Embedding Overlapping Protocol
10:14

3D Scanning Technology Bridging Microcircuits and Macroscale Brain Images in 3D Novel Embedding Overlapping Protocol

Published on: May 12, 2019

7.3K

向更少受约束的宏观神经架构搜索

Vasco Lopes, Luis A Alexandre

    IEEE transactions on neural networks and learning systems
    |October 31, 2023
    PubMed
    概括

    较少受约束的宏神经架构搜索 (LCMNAS) 能够自动发现高性能神经网络. 这种方法探索更广泛的搜索空间没有人类启发式,有效地实现最先进的结果.

    科学领域:

    • 人工智能的人工智能
    • 机器学习 机器学习
    • 计算机视觉 计算机视觉

    背景情况:

    • 神经架构搜索 (NAS) 方法通常依赖于人类定义的约束,限制了对新型网络架构的探索.
    • 现有的NAS方法通常专注于基于单元的搜索空间,限制整个网络结构的设计 (宏观搜索).

    研究的目的:

    • 为更广泛的建筑探索引入更少受约束的宏观神经架构搜索 (LCMNAS).
    • 开发一种NAS方法,在没有预定义的启发式或边界搜索空间的情况下执行宏观搜索.
    • 为了实现最先进的性能,降低计算成本.

    主要方法:

    • LCMNAS使用加权定向图 (WDGs) 来自主生成复杂的,不那么受限制的搜索空间,由现有架构提供信息.
    • 一个进化的搜索策略从头开始生成完整的架构.
    • 混合性能估计方法将初始架构信息与低准确度估计相结合,以预测性能.

    主要成果:

    • 在14个不同的数据集中,LCMNAS成功生成了基于细胞和基于宏的架构.
    • 该方法以最小的GPU计算实现了最先进的结果.
    • 广泛的研究证实了LCMNAS组件在细胞和宏观搜索设置中的有效性.

    更多相关视频

    Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology
    09:44

    Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology

    Published on: March 8, 2024

    4.8K
    Anatomically Inspired Three-dimensional Micro-tissue Engineered Neural Networks for Nervous System Reconstruction, Modulation, and Modeling
    10:45

    Anatomically Inspired Three-dimensional Micro-tissue Engineered Neural Networks for Nervous System Reconstruction, Modulation, and Modeling

    Published on: May 31, 2017

    13.1K

    相关实验视频

    Last Updated: Jul 12, 2025

    3D Scanning Technology Bridging Microcircuits and Macroscale Brain Images in 3D Novel Embedding Overlapping Protocol
    10:14

    3D Scanning Technology Bridging Microcircuits and Macroscale Brain Images in 3D Novel Embedding Overlapping Protocol

    Published on: May 12, 2019

    7.3K
    Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology
    09:44

    Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology

    Published on: March 8, 2024

    4.8K
    Anatomically Inspired Three-dimensional Micro-tissue Engineered Neural Networks for Nervous System Reconstruction, Modulation, and Modeling
    10:45

    Anatomically Inspired Three-dimensional Micro-tissue Engineered Neural Networks for Nervous System Reconstruction, Modulation, and Modeling

    Published on: May 31, 2017

    13.1K

    结论:

    • 通过实现不受约束的宏观搜索,LCMNAS显著提升了NAS,从而导致了新的高性能神经网络架构.
    • 提出的方法提供了一种高效和自动化的方法来发现复杂的架构,优于人类设计的网络.