相关实验视频

Updated: May 8, 2026

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

Published on: February 23, 2017

BiasPruner:在持续学习中缓解偏差转移,以进行公平的医学图像分析

Nourhan Bayasi¹, Jamil Fayyad², Alceu Bissoto³

¹University of British Columbia, Vancouver, BC, Canada.

Medical image analysis

|August 20, 2025

概括

持续学习 (CL) 方法扩大了偏见,危及人工智能的公平性. BiasPruner通过修剪有偏见的网络单元来缓解这一问题,提高医学成像AI的准确性和公平性.

科学领域:

人工智能
机器学习
医学成像

背景情况:

持续学习 (CL) 允许模型顺序学习,但往往会转移偏差.
偏差转移会对人工智能的公平性和性能产生负面影响,尤其是在关键的医疗应用中.
现有的CL方法可以放大虚假的相关性,导致不平等的治疗和错误诊断.

研究的目的:

提出一个新的框架,BiasPruner,以缓解持续学习中的偏见传播.
提高人工智能模型的公平性和准确性,特别是在医学成像任务中.
通过传统的CL方法来解决偏差的扩大.

主要方法:

BiasPruner使用偏差归因得分来识别和削减对偏差负责的网络单元.
为每项任务创建基于数据的子网络, 保存先前的知识并防止灾难性的遗忘.
在推断过程中使用任务不可知门机制进行强有力的预测.

主要成果:

BiasPruner在精度和公平性方面显著超过了最先进的CL方法.
在皮肤病变和胸部X射线分类等医学成像任务中有效缓解偏差.
在保持顺序学习能力的同时成功阻止了偏差转移.

结论:

BiasPruner提供了一个强大的解决方案,
该框架提高了人工智能公平性和可靠性, 这对于公平的医疗保健至关重要.
在开发可信且公正的人工智能系统方面,

关键词:

偏差转移灾难性的遗忘持续学习没有偏见的表述脱偏差子网络动态子网络修剪虚假的相关性

更多相关视频

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

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

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

Published on: November 30, 2022

相关概念视频

Improving Translational Accuracy

Improving Translational Accuracy

Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...

Improving Translational Accuracy

Improving Translational Accuracy

Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...