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: Dec 23, 2025

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
08:20

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images

Published on: October 27, 2023

2.3K

DeepDetectNet vs RLAttackNet: An adversarial method to improve deep learning-based static malware detection model.

Yong Fang1, Yuetian Zeng1, Beibei Li1

  • 1College of Cybersecurity, Sichuan University, Chengdu, China.

Plos One
|April 24, 2020
PubMed
Summary
This summary is machine-generated.

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

DDML: Multi-Student Knowledge Distillation for Hate Speech.

Entropy (Basel, Switzerland)·2025
Same author

[Changes in portal vein and hepatic vein blood flow volume and their ratio in SD rats during induced carcinogenesis of hepatocellular carcinoma].

Nan fang yi ke da xue xue bao = Journal of Southern Medical University·2015
Same author

Modeling the relationship of epigenetic modifications to transcription factor binding.

Nucleic acids research·2015
Same author

Artesunate induces apoptosis and inhibits growth of Eca109 and Ec9706 human esophageal cancer cell lines in vitro and in vivo.

Molecular medicine reports·2015
Same author

Chronopharmacodynamics and mechanisms of antitumor effect induced by erlotinib in xenograft-bearing nude mice.

Biochemical and biophysical research communications·2015
Same author

Characterization of protein alterations in damaged axons in the brainstem following traumatic brain injury using fourier transform infrared microspectroscopy: a preliminary study.

Journal of forensic sciences·2015
Same journal

Modeling and analysis of forward and inverse kinematics for a flexible Stewart platform.

PloS one·2026
Same journal

Barriers and facilitators to healthcare utilization amongst people living with sickle cell disease in the United States: A scoping review.

PloS one·2026
Same journal

Enhancing data completeness in time series: Imputation strategies for missing data using significant periodically correlated components.

PloS one·2026
Same journal

Key targets and mechanisms by which gut microbiota-derived metabolites regulate Alzheimer's disease through the immune - inflammatory pathway: Based on network pharmacology and molecular docking.

PloS one·2026
Same journal

Grid-tied Transformer-less Boost Switched Capacitor Topology (TLBSCT) for PV applications.

PloS one·2026
Same journal

The load-velocity profiles and exercise-specific velocity zones for seven commonly used weightlifting exercises.

PloS one·2026
See all related articles

This study introduces an automated method using deep reinforcement learning to generate adversarial malware samples, significantly improving static malware detection model robustness and reducing attack success rates.

Area of Science:

  • Cybersecurity
  • Machine Learning
  • Software Engineering

Background:

  • Deep learning enhances static malware detection by identifying novel threats.
  • Deep learning models are susceptible to adversarial attacks, particularly gradient-based attacks on Windows PE file features.

Purpose of the Study:

  • To address limitations in existing adversarial sample generation methods (manual control, low efficiency).
  • To propose a novel, automated adversarial sample generation method using deep reinforcement learning for static malware detection.

Main Methods:

  • Developed DeepDetectNet, a deep learning-based static Portable Executable (PE) malware detection model (initial AUC 0.989).
  • Implemented RLAttackNet, a reinforcement learning model for generating adversarial malware samples designed to evade DeepDetectNet.

More Related Videos

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

931

Related Experiment Videos

Last Updated: Dec 23, 2025

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
08:20

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images

Published on: October 27, 2023

2.3K
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

931
  • Utilized generated adversarial samples to retrain and reinforce the DeepDetectNet model.

    • Main Results:

    • RLAttackNet successfully generated malware samples that bypassed DeepDetectNet, achieving an attack success rate of 19.13%.
    • Retraining DeepDetectNet with these adversarial samples improved its AUC from 0.989 to 0.996.
    • The attack success rate against the retrained model dropped significantly from 19.13% to 3.1%.

    Conclusions:

    • The proposed RLAttackNet offers an efficient and universal method for automatic adversarial sample generation in static malware detection.
    • Adversarial training with samples generated by RLAttackNet substantially enhances the resilience and performance of deep learning-based malware detectors.