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 Experiment Video

Updated: Jun 27, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Residual-guided hybrid framework for adversarially robust deep learning-based network intrusion detection.

Sudip Saha1, Muhammad Arslan Pervaiz1, Muhammad Safwat Rahman2

  • 1Department of Cybersecurity, Pace University, New York, New York, United States of America.

Plos One
|June 1, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Multi-modal federated learning with differential privacy for privacy-preserving healthcare AI.

Scientific reports·2026
Same author

Evaluating the Impact of an Emotion-Focused Model of Dementia Care on Patient Outcomes in Acute Care Settings.

Canadian geriatrics journal : CGJ·2026
Same author

Acute Respiratory Distress in a Hospitalized Patient Revealing Cervical Diffuse Idiopathic Skeletal Hyperostosis (DISH).

Cureus·2025
Same author

Portal vein thrombosis as a complication of Eltrombopag used in the treatment of immune thrombocytopenia in a patient of chronic liver disease.

Indian journal of pharmacology·2025
Same author

Retraction notice to "Interaction of AtHMGB15, an ARID-HMG family protein, with RING-H2 type E3 ubiquitin ligase AtATL79" [Plant Sci. 276 (2018) 32-43].

Plant science : an international journal of experimental plant biology·2024
Same author

A universal bacterial sensor created by integrating a light modulating aptamer complex with photoelectrochemical signal readout.

Biosensors & bioelectronics·2023
This summary is machine-generated.

This study introduces a new deep learning framework enhancing machine learning model security against cyber threats. The hybrid model improves adversarial robustness and accuracy, crucial for sensitive applications.

Area of Science:

  • Cybersecurity
  • Machine Learning
  • Deep Learning

Background:

  • Traditional deep learning models are vulnerable to adversarial attacks, impacting sensitive sectors like healthcare and finance.
  • Cyber threats are increasingly sophisticated, challenging the reliability of machine learning in real-world systems.

Purpose of the Study:

  • To develop a novel hybrid adversarially-trained deep learning framework.
  • To enhance resilience against Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) attacks.
  • To optimize both clean accuracy and adversarial robustness.

Main Methods:

  • Integration of reinforcement learning-inspired robustness adaptation.
  • Incorporation of knowledge-driven regularization.
  • Simultaneous optimization of cross-entropy and adversarial loss, monitoring calibration error and gradient dynamics.

Related Experiment Videos

Last Updated: Jun 27, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Main Results:

  • Achieved up to 97.88% accuracy on clean data.
  • Maintained 84.9% accuracy under FGSM and 81.75% under PGD attacks.
  • Outperformed CNN and LSTM baselines by 6-10 percentage points in adversarial robustness.
  • Demonstrated a 30% reduction in expected calibration error.

Conclusions:

  • The proposed framework offers improved adversarial robustness and stable convergence.
  • The model shows scalability and efficiency, validating its practical application.
  • This work provides insights into balancing robustness, efficiency, and generalization in machine learning.