skip to main content
訪客
個人書架
我的帳戶
登出
登入
This feature requires javascript
檢索首頁
圖書館首頁
電子期刊
引用參考文獻查詢
指定參考書查詢
新書通報
標籤查詢
線上輔助
語言:
English
繁體中文
This feature required javascript
This feature requires javascript
Primo Search
館藏+文章
館藏+文章
館藏
查館藏
文章
查文章
機構典藏
機構典藏
Search For:
Clear Search Box
Search in:
文章
Or hit Enter to replace search target
Or select another collection:
Search in:
文章
進階檢索
瀏覽查詢
This feature requires javascript
顯示結果:
查詢種類
criteria input
包含在我的檢索語句內
完全相同
顯示結果:
查詢種類 索引
criteria input
任何地方
題名
ISBN
ISSN
Show Results with:
題名
Show Results with:
任何地方
題名
ISBN
ISSN
This feature requires javascript
Training confounder-free deep learning models for medical
applications
Zhao, Qingyu ; Adeli, Ehsan ; Pohl, Kilian M
Nature communications, 2020-11, Vol.11 (1), p.6010-6010, Article 6010
[同儕審閱期刊]
可取得全文
引用
被引用
線上檢視
詳細格式
評論和標籤
相關文章推薦
FullText@NUTN
引用次數
This feature requires javascript
傳送到
加入個人書架
從個人書架中移除
E-mail
列印
永久連結
引用
EndNote
導出 RiS
This feature requires javascript
題名:
Training confounder-free deep learning models for medical
applications
著者:
Zhao, Qingyu
;
Adeli, Ehsan
;
Pohl, Kilian M
主題:
Adolescent
;
Adolescents
;
Adult
;
Age Determination by Skeleton - methods
;
Age Factors
;
Aged
;
Aged, 80 and over
;
Brain - diagnostic imaging
;
Case-Control Studies
;
Child
;
Child development
;
Chronology
;
Cohort Studies
;
Consortia
;
Datasets as Topic
;
Deep Learning
;
Diagnosis
;
Feature extraction
;
Female
;
Gender aspects
;
Hand - diagnostic imaging
;
HIV
;
HIV Infections - diagnosis
;
Human immunodeficiency virus
;
Humans
;
Image Interpretation, Computer-Assisted - methods
;
Invariants
;
Learning algorithms
;
Machine learning
;
Magnetic resonance
;
Magnetic Resonance Imaging
;
Male
;
Mathematical models
;
Medical imaging
;
Middle Aged
;
Neuroimaging
;
Prediction models
;
Predictive Value of Tests
;
Radiography
;
Sex Characteristics
;
Sex differences
;
Sex Factors
;
Sexually transmitted diseases
;
Statistical analysis
;
Statistical methods
;
Statistical models
;
STD
;
Training
;
Young Adult
所屬期刊:
Nature communications, 2020-11, Vol.11 (1), p.6010-6010, Article 6010
描述:
The presence of confounding effects (or biases) is one of the most critical challenges in using deep learning to advance discovery in medical imaging studies. Confounders affect the relationship between input data (e.g., brain MRIs) and output variables (e.g., diagnosis). Improper modeling of those relationships often results in spurious and biased associations. Traditional machine learning and statistical models minimize the impact of confounders by, for example, matching data sets, stratifying data, or residualizing imaging measurements. Alternative strategies are needed for state-of-the-art deep learning models that use end-to-end training to automatically extract informative features from large set of images. In this article, we introduce an end-to-end approach for deriving features invariant to confounding factors while accounting for intrinsic correlations between the confounder(s) and prediction outcome. The method does so by exploiting
concepts
from traditional statistical methods and recent fair machine learning schemes. We evaluate the method on predicting the diagnosis of HIV solely from Magnetic Resonance Images (MRIs), identifying morphological sex differences in adolescence from those of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), and determining the bone age from X-ray images of children. The results show that our method can accurately predict while reducing biases associated with confounders. The code is available at https://github.com/qingyuzhao/br-net .
出版者:
England: Nature Publishing Group
語言:
英文
識別號:
ISSN: 2041-1723
EISSN: 2041-1723
DOI: 10.1038/s41467-020-19784-9
PMID: 33243992
資源來源:
Publicly Available Content Database
DOAJ Directory of Open Access Journals
連結
View this record in MEDLINE/PubMed
This feature requires javascript
This feature requires javascript
返回到檢索清單
前一個記錄
檢索清單
10
下ㄧ個記錄
This feature requires javascript
This feature requires javascript
正在檢索遠程資料庫,請稍等
查詢:
在
primo_central_multiple_fe
顯示現有記錄
This feature requires javascript
This feature requires javascript