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

Convolution: Math, Graphics, and Discrete Signals01:24

Convolution: Math, Graphics, and Discrete Signals

504
In any LTI (Linear Time-Invariant) system, the convolution of two signals is denoted using a convolution operator, assuming all initial conditions are zero. The convolution integral can be divided into two parts: the zero-input or natural response and the zero-state or forced response, with t0 indicating the initial time.
To simplify the convolution integral, it is assumed that both the input signal and impulse response are zero for negative time values. The graphical convolution process...
504
Neural Circuits01:25

Neural Circuits

1.9K
Neural circuits and neuronal pools are two of the main structures found in the nervous system. Neural circuits are networks of neurons that work together to carry out a specific task or process. They consist of interconnected neurons and glial cells, which provide structural and metabolic support.
Neuronal pools are collections of nerve cells with similar functions and interact through chemical and electrical signals. These pools include both interneurons (the central neural circuit nodes that...
1.9K
Convolution Properties I01:20

Convolution Properties I

294
Convolution computations can be simplified by utilizing their inherent properties.
The commutative property reveals that the input and the impulse response of an LTI (Linear Time-Invariant) system can be interchanged without affecting the output:
294
Improving Translational Accuracy02:07

Improving Translational Accuracy

12.0K
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...
12.0K
Convolution Properties II01:17

Convolution Properties II

332
The important convolution properties include width, area, differentiation, and integration properties.
The width property indicates that if the durations of input signals are T1 and T2, then the width of the output response equals the sum of both durations, irrespective of the shapes of the two functions. For instance, convolving two rectangular pulses with durations of 2 seconds and 1 second results in a function with a width of 3 seconds.
The area property asserts that the area under the...
332
Reducing Line Loss01:18

Reducing Line Loss

222
In a three-phase circuit, line loss is an indicator of energy dissipated as heat due to the resistance of transmission lines. To address this, incorporating transformers into the system—a step-up transformer at the source and a step-down transformer at the load—is a strategic solution. Two three-phase transformers are introduced to improve this.
With a step-up transformer at the source, the voltage is increased, thereby reducing the current in the transmission lines since power loss...
222

You might also read

Related Articles

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

Sort by
Same author

[Etiology and pathogens of fungal endophthalmitis].

[Zhonghua yan ke za zhi] Chinese journal of ophthalmology·2015
Same author

Eucommia ulmoides Oliv. bark aqueous extract inhibits osteoarthritis in a rat model of osteoarthritis.

Journal of ethnopharmacology·2015
Same author

Aucubin prevents interleukin-1 beta induced inflammation and cartilage matrix degradation via inhibition of NF-κB signaling pathway in rat articular chondrocytes.

International immunopharmacology·2015
Same author

Treatment with recombinant lubricin attenuates osteoarthritis by positive feedback loop between articular cartilage and subchondral bone in ovariectomized rats.

Bone·2015
Same author

Tet1-mediated DNA demethylation regulates neuronal cell death induced by oxidative stress.

Scientific reports·2015
Same author

Authors' reply.

Arthroscopy : the journal of arthroscopic & related surgery : official publication of the Arthroscopy Association of North America and the International Arthroscopy Association·2015

Related Experiment Video

Updated: Oct 20, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
09:47

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

1.4K

Structural Compression of Convolutional Neural Networks with Applications in Interpretability.

Reza Abbasi-Asl1,2,3, Bin Yu3,4

  • 1Department of Neurology, Department of Bioengineering and Therapeutic Sciences, University of California, San Francisco, San Francisco, CA, United States.

Frontiers in Big Data
|September 13, 2021
PubMed
Summary

This study introduces a method to compress deep convolutional neural networks (CNNs), making them smaller and more interpretable. The technique prunes filters based on their contribution to accuracy, enhancing scientific understanding of these complex models.

Keywords:
compressionconvolutional neural networksfilter diversityfilter pruninginterpretation

Related Experiment Videos

Last Updated: Oct 20, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
09:47

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

1.4K

Area of Science:

  • Computer Vision
  • Machine Learning
  • Artificial Intelligence

Background:

  • Deep convolutional neural networks (CNNs) excel in machine vision but their complexity, with millions of weights and thousands of filters, hinders scientific interpretation.
  • Understanding the function of individual filters within CNNs is crucial for scientific insight and trust.

Purpose of the Study:

  • To develop a greedy structural compression scheme for creating smaller and more interpretable CNNs.
  • To maintain close to original classification accuracy after compression.
  • To demonstrate the interpretability of the compressed CNNs.

Main Methods:

  • A greedy structural compression scheme based on pruning filters.
  • Utilizing the Classification Accuracy Reduction (CAR) importance index to identify filters with minimal contribution.
  • Introducing a variant of CAR to quantify the importance of image categories to each filter.

Main Results:

  • The compression scheme successfully reduced the number of filters by an order of magnitude while preserving accuracy.
  • CAR-compressed CNNs demonstrated enhanced interpretability, pruning redundant filters like color filters.
  • The CAR variant revealed meaningful interpretations of filters by linking them to specific image categories.

Conclusions:

  • The proposed compression method yields smaller, more interpretable CNNs suitable for scientific applications.
  • Pruning filters using the CAR index effectively removes redundancy and aids in understanding CNN functionalities.
  • The enhanced interpretability facilitates the analysis of filter-category relationships in deep learning models.