Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Overview of Nitrogen Metabolism01:20

Overview of Nitrogen Metabolism

9.5K
Nitrogen is a very important element for life because it is a major constituent of proteins and nucleic acids. It is a macronutrient, and in nature, it is recycled from organic compounds and stored in the form of  ammonia, ammonium ions, nitrate, nitrite, or  nitrogen gas by many metabolic processes. Many of these metabolic processes are carried out only by prokaryotes.
The largest pool of nitrogen available in the terrestrial ecosystem is gaseous nitrogen (N2) from the air, but this...
9.5K
Impact of Schemas01:30

Impact of Schemas

27
Schemas are cognitive structures that provide a framework for interpreting and organizing social information. They help individuals navigate complex environments by offering expectations about people, events, and behaviors. Schemas influence attention, encoding, and retrieval processes, thereby shaping the entire trajectory of information processing in social contexts.Attention and Cognitive LoadDuring initial attention, schemas function as filters that prioritize schema-consistent information,...
27
Mass Analyzers: Common Types01:19

Mass Analyzers: Common Types

880
The quadrupole mass analyzer consists of four cylindrical metal rods arranged in a diamond carrying a DC voltage and a radio-frequency AC voltage. The motion of ions through the quadrupole depends on the field strength, causing only ions of a certain m/z to resonate successfully and strike the detector at a given field strength. Though the transmission rate for these analyzers is high, the exact elemental composition of the sample is not determined because of low resolution; however, they are...
880
Storage01:23

Storage

156
A schema is a mental framework that helps individuals organize and interpret information. Schemata, formed from previous experiences, influence how we process new information: how we encode it, the inferences we make, and how we retrieve it. For instance, a schema for what a typical classroom looks like might include desks, a teacher's desk, a whiteboard, and students in such an environment. This expectation helps us quickly understand and navigate new classrooms without needing to analyze...
156
ER Retrieval Pathway01:45

ER Retrieval Pathway

4.0K
In the secretory pathway, vesicles transport proteins from one cellular compartment to another in forward transport to deliver the protein to its correct location. Occasionally, misfolded proteins and incorrect proteins escape their original compartments, and a retrieval pathway is used to return the escaped proteins to their original compartment.
The ER uses many checkpoints to prevent the entry of incorrectly folded or a resident protein as cargo onto a transport vesicle. These mechanisms...
4.0K
Microbial Classification System01:24

Microbial Classification System

299
Classification is the process of organizing organisms into hierarchically inclusive groups based on their phenotypic similarities or evolutionary relationships. A species comprises one or more strains, and closely related species are grouped into genera. Genera are further classified into families, families into orders, orders into classes, and so forth, up to the domain level, which is the broadest taxonomic rank derived from a combination of phenotypic and genotypic data.The nomenclature of...
299

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Leveraging Machine Learning to Advance Alcohol Research: Current Applications, Challenges, and Opportunities.

Alcohol research : current reviews·2026
Same author

A framework of digital biomarkers for neurodegenerative diseases.

Nature reviews bioengineering·2026
Same author

SocialGen: Modeling Multi-Human Social Interaction with Language Models.

Proceedings. International Conference on 3D Vision·2026
Same author

Mapping Individualized Developmental Imbalance in Youth and Its Association with Psychopathology.

bioRxiv : the preprint server for biology·2026
Same author

Using deep learning to identify brain networks mediating cognitive and motor impairments in alcohol use disorder.

Translational psychiatry·2026
Same author

A generalized synthetic control algorithm for sparse functional data.

bioRxiv : the preprint server for biology·2026
Same journal

CARL: A Framework for Equivariant Image Registration.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
Same journal

Unifying Top-down and Bottom-up Scanpath Prediction Using Transformers.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
Same journal

Latent Drifting in Diffusion Models for Counterfactual Medical Image Synthesis.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
Same journal

The Language of Motion: Unifying Verbal and Non-verbal Language of 3D Human Motion.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
Same journal

Perceptual Inductive Bias Is What You Need Before Contrastive Learning.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
Same journal

MultiMorph: On-demand Atlas Construction.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition·2026
See all related articles

Related Experiment Video

Updated: Oct 13, 2025

Executing Complexity-Increasing Queries in Relational MySQL and NoSQL MongoDB and EXist Size-Growing ISO/EN 13606 Standardized EHR Databases
07:26

Executing Complexity-Increasing Queries in Relational MySQL and NoSQL MongoDB and EXist Size-Growing ISO/EN 13606 Standardized EHR Databases

Published on: March 19, 2018

9.5K

Metadata Normalization.

Mandy Lu1, Qingyu Zhao1, Jiequan Zhang1

  • 1Stanford University, Stanford, CA 94305.

Proceedings. IEEE Computer Society Conference on Computer Vision and Pattern Recognition
|November 15, 2021
PubMed
Summary
This summary is machine-generated.

Metadata Normalization (MDN) is a new deep learning layer that removes bias from extraneous variables. This method corrects feature distributions, improving model accuracy across diverse datasets.

More Related Videos

Cloud-Based Phrase Mining and Analysis of User-Defined Phrase-Category Association in Biomedical Publications
09:20

Cloud-Based Phrase Mining and Analysis of User-Defined Phrase-Category Association in Biomedical Publications

Published on: February 23, 2019

8.9K
A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
07:50

A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts

Published on: September 20, 2018

16.1K

Related Experiment Videos

Last Updated: Oct 13, 2025

Executing Complexity-Increasing Queries in Relational MySQL and NoSQL MongoDB and EXist Size-Growing ISO/EN 13606 Standardized EHR Databases
07:26

Executing Complexity-Increasing Queries in Relational MySQL and NoSQL MongoDB and EXist Size-Growing ISO/EN 13606 Standardized EHR Databases

Published on: March 19, 2018

9.5K
Cloud-Based Phrase Mining and Analysis of User-Defined Phrase-Category Association in Biomedical Publications
09:20

Cloud-Based Phrase Mining and Analysis of User-Defined Phrase-Category Association in Biomedical Publications

Published on: February 23, 2019

8.9K
A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
07:50

A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts

Published on: September 20, 2018

16.1K

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Deep Learning

Background:

  • Batch Normalization (BN) techniques normalize feature distributions using batch statistics to mitigate covariate shift in deep learning.
  • However, BN does not address biases introduced by extraneous variables or multiple data distributions, known as metadata.
  • Metadata can induce confounding effects, impacting model fairness and performance, such as race influencing gender classification.

Purpose of the Study:

  • To introduce Metadata Normalization (MDN), a novel layer designed to correct the influence of metadata on feature distributions within deep learning models.
  • To develop an end-to-end trainable operation that effectively removes metadata-induced bias during the training process.
  • To provide a method for quantifying and mitigating confounding effects caused by metadata in machine learning.

Main Methods:

  • The Metadata Normalization (MDN) layer is proposed as a batch-level operation integrated into deep learning frameworks.
  • MDN employs regression analysis techniques to "regress out" the effects of metadata from model features during training.
  • Distance correlation is used as a metric to quantify the distribution bias attributable to metadata.

Main Results:

  • The MDN layer successfully removes the influence of metadata on feature distributions across diverse datasets.
  • Demonstrated effectiveness in four distinct settings: synthetic data, 2D image classification, video analysis, and 3D medical imaging.
  • Validation confirms the method's capability to mitigate bias introduced by confounding variables.

Conclusions:

  • Metadata Normalization (MDN) offers a robust solution for addressing metadata-induced bias in deep learning.
  • The MDN layer can be seamlessly incorporated into existing training pipelines to enhance model fairness and reliability.
  • This approach significantly improves the generalizability and trustworthiness of deep learning models by correcting for confounding factors.