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

Woodward–Hoffmann Selection Rules and Microscopic Reversibility01:34

Woodward–Hoffmann Selection Rules and Microscopic Reversibility

3.6K
Electrocyclic reactions, cycloadditions, and sigmatropic rearrangements are concerted pericyclic reactions that proceed via a cyclic transition state. These reactions are stereospecific and regioselective. The stereochemistry of the products depends on the symmetry characteristics of the interacting orbitals and the reaction conditions. Accordingly, pericyclic reactions are classified as either symmetry-allowed or symmetry-forbidden. Woodward and Hoffmann presented the selection criteria for...
3.6K
Decision Making: P-value Method01:09

Decision Making: P-value Method

6.6K
The process of hypothesis testing based on the P-value method includes calculating the P- value using the sample data and interpreting it.
First, a specific claim about the population parameter is proposed. The claim is based on the research question and is stated in a simple form. Further, an opposing statement to the claim  is also stated. These statements can act as null and alternative hypotheses:  a null hypothesis would be a neutral statement while the alternative hypothesis can...
6.6K
Network Function of a Circuit01:25

Network Function of a Circuit

507
Frequency response analysis in electrical circuits provides vital insights into a circuit's behavior as the frequency of the input signal changes. The transfer function, a mathematical tool, is instrumental in understanding this behavior. It defines the relationship between phasor output and input and comes in four types: voltage gain, current gain, transfer impedance, and transfer admittance. The critical components of the transfer function are the poles and zeros.
507
Reversible and Irreversible Processes01:14

Reversible and Irreversible Processes

5.3K
The thermodynamic processes can be classified into reversible and irreversible processes. The processes that can be restored to their initial state are called reversible processes. It is only possible if the process is in quasi-static equilibrium, i.e., it takes place in infinitesimally small steps, and the system remains at equilibrium However, these are ideal processes and do not occur naturally. An ideal system undergoing a reversible process is always in thermodynamic equilibrium within...
5.3K
Decision Making01:20

Decision Making

562
Decision-making is a fundamental cognitive process that involves evaluating alternatives and selecting among them. This process can range from simple choices, such as deciding what to wear, to complex decisions, like choosing a major in college or a career path. The complexity of the decision often dictates the approach we use, which can be broadly categorized into two types: automatic and controlled decision-making.
Automatic decision-making is fast, intuitive, and relies on gut feelings...
562
Thevinin's Theorem01:15

Thevinin's Theorem

1.2K
Thévenin's theorem plays a pivotal role in electrical circuit analysis, offering a solution to the challenges posed by variable loads within a circuit. In practical applications, it is common to encounter circuits where certain elements remain fixed while others fluctuate, often referred to as the "load." A typical household electrical outlet serves as a prime example of a variable load, as it can be connected to a variety of appliances, each with its own unique electrical characteristics.
1.2K

You might also read

Related Articles

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

Sort by
Same author

Unveiling the synergy of Interface and defects in V<sub>2</sub>O<sub>3</sub>/VN heterojunctions for high-performance aqueous zinc-ion batteries.

Journal of colloid and interface science·2026
Same author

Machine learning modeling to predict HCC locations in cirrhotic patients undergoing MRI - a proof-of-concept study.

Abdominal radiology (New York)·2026
Same author

Bioinspired Rhombic VO<sub>2</sub> Metasurface with Low Solar Absorptance for Self-adaptive All-Weather Building Thermal Management.

ACS applied materials & interfaces·2026
Same author

School screening for adolescent idiopathic scoliosis in China: a five-year evaluating the evaluation in Zhongshan city.

Scientific reports·2026
Same author

3-Monochloro-1,2-propanediol disrupts ovarian germline stem cell maintenance via Golgi stress-induced autophagy in Drosophila.

Chemico-biological interactions·2026
Same author

Intraoperative MEP and EMG during the wake-up phase as adjuncts to the wake-up test in severe spinal deformity surgery.

Frontiers in surgery·2026

Related Experiment Video

Updated: Dec 6, 2025

Inherent Dynamics Visualizer, an Interactive Application for Evaluating and Visualizing Outputs from a Gene Regulatory Network Inference Pipeline
10:44

Inherent Dynamics Visualizer, an Interactive Application for Evaluating and Visualizing Outputs from a Gene Regulatory Network Inference Pipeline

Published on: December 7, 2021

2.5K

Decision Explanation and Feature Importance for Invertible Networks.

Juntang Zhuang1, Nicha C Dvornek2, Xiaoxiao Li1

  • 1Biomedical Engineering, Yale University, New Haven, CT USA.

... IEEE International Conference on Computer Vision Workshops. IEEE International Conference on Computer Vision
|October 7, 2020
PubMed
Summary

Invertible neural networks offer insights into black-box models. This method explains model decisions by inverting decision boundaries and analyzing feature importance for better understanding deep learning.

More Related Videos

Creating Objects and Object Categories for Studying Perception and Perceptual Learning
14:38

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Published on: November 2, 2012

12.1K

Related Experiment Videos

Last Updated: Dec 6, 2025

Inherent Dynamics Visualizer, an Interactive Application for Evaluating and Visualizing Outputs from a Gene Regulatory Network Inference Pipeline
10:44

Inherent Dynamics Visualizer, an Interactive Application for Evaluating and Visualizing Outputs from a Gene Regulatory Network Inference Pipeline

Published on: December 7, 2021

2.5K
Creating Objects and Object Categories for Studying Perception and Perceptual Learning
14:38

Creating Objects and Object Categories for Studying Perception and Perceptual Learning

Published on: November 2, 2012

12.1K

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computer Vision

Background:

  • Deep neural networks (DNNs) are often considered black boxes, making them vulnerable to adversarial attacks and difficult to interpret.
  • The inherent complexity and non-linear transformations within DNNs obscure the reasoning behind their predictions.

Purpose of the Study:

  • To develop methods for interpreting and explaining the behavior of deep neural networks, particularly addressing their black-box nature.
  • To leverage invertible neural networks for enhanced model interpretability and to provide explanations for adversarial robustness.

Main Methods:

  • Utilizing invertible neural networks to accurately reconstruct layer inputs from outputs, enabling a two-stage classification process (invertible transformation followed by linear classification).
  • Inverting the decision boundary from the feature space back to the input space to visualize and understand classification boundaries.
  • Defining explanations as the difference between a data point and its projection onto the decision boundary.
  • Employing first-order Taylor expansion for local approximation of neural networks and defining feature importance via a local linear model.

Main Results:

  • Demonstrated the potential of invertible networks to unravel the black-box nature of DNNs by enabling input-output reconstruction.
  • Successfully inverted decision boundaries to the input space, providing a visualizable explanation of classifier behavior.
  • Proposed a novel definition of explanation based on data point projection and introduced a method for calculating feature importance using local linear models.

Conclusions:

  • Invertible neural networks provide a powerful framework for enhancing the interpretability of deep learning models.
  • The proposed explanation methods offer insights into model predictions and feature importance, contributing to more transparent and trustworthy AI systems.
  • The developed techniques can aid in understanding and potentially mitigating adversarial attacks by clarifying model decision-making processes.