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

Prevalence and Burden of Cardiac Amyloidosis in a Population-Based Autopsy Cohort.

Circulation·2026
Same author

Establishment of harmonized international reference ranges for plasma estradiol concentrations in postmenopausal women.

The Journal of clinical endocrinology and metabolism·2026
Same author

Performance of clinical breast cancer risk prediction models versus a mammography-based artificial intelligence risk model.

Journal of the National Cancer Institute·2026
Same author

Circulating levels of insulin-like growth factor I (IGF-I) and risk of multiple myeloma: An observational and Mendelian randomisation study.

British journal of haematology·2026
Same author

Multimodal imaging reveals a lysosomal drug reservoir that drives heterogeneous distribution of PARP inhibitors.

Nature communications·2026
Same author

ULK4 and CDKN2A polymorphisms influence the risk of developing monoclonal gammopathy of undetermined significance.

International journal of cancer·2026

相关实验视频

Updated: Jul 9, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K

使用概念激活矢量进行高效的对抗性脱皮 - - 医学图像案例研究

Ramon Correa1, Khushbu Pahwa2, Bhavik Patel3

  • 1Arizona State University, SCAI, Tempe, AZ, 85281, USA.

Journal of biomedical informatics
|December 3, 2023
PubMed
概括
此摘要是机器生成的。

我们开发了一种使用概念激活矢量 (CAV) 的对抗性去偏差方法,以减少AI模型在医学成像中的偏差. 这种方法提高了未见的群体的性能,超过了标准的微调策略.

关键词:
对抗性的公平和公平.概念激活向量的概念激活向量调整偏差 调整偏差乳房造影图像 乳房造影图像一些X射线图像.

更多相关视频

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

1.5K
Basics of Multivariate Analysis in Neuroimaging Data
06:35

Basics of Multivariate Analysis in Neuroimaging Data

Published on: July 24, 2010

16.9K

相关实验视频

Last Updated: Jul 9, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K
Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

1.5K
Basics of Multivariate Analysis in Neuroimaging Data
06:35

Basics of Multivariate Analysis in Neuroimaging Data

Published on: July 24, 2010

16.9K

科学领域:

  • 人工智能的人工智能
  • 医疗成像医学成像
  • 机器学习偏见 机器学习偏见

背景情况:

  • 人工智能模型在实时部署时面临挑战,原因是未见的人群的可信性问题.
  • 复杂的人工智能模型经常充当黑子,在决策中表现出隐含的偏见,特别是在少数群体的子组中.

研究的目的:

  • 为人工智能模型开发一种高效的对抗性去偏差方法.
  • 减少人工智能决策中的种族差异,同时保持任务性能.
  • 适应模型解释性技术以减轻偏差.

主要方法:

  • 开发了一种使用概念激活矢量 (CAV) 的对抗性去偏差方法.
  • 利用CAV来识别负责学习种族的卷积层.
  • 通过微调仅确定层的应用部分学习,最大限度地减少性能下降.

主要成果:

  • 在胸部X射线和乳房造影数据集上进行评估,并进行外部验证.
  • 偏差的胸部X射线模型实现了比基线 (0.57) 和微调模型 (0.81) 更高的AUC (0.87).
  • 乳房影像模型在调试后在外部数据集 (AUC 0.8 到 0.86) 上提高了性能.

结论:

  • 内部训练的对抗模型的性能与外部数据的标准微调相比或更好.
  • 对抗式训练方法是基于梯度的模型的模型架构不可知.
  • 培训代码是在学术开源许可证下提供的.