Scribd
Upload
24 views10 pages

GR20 Final

Uploaded by

pentasanjananaidu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
24 views10 pages

GR20 Final

Uploaded by

pentasanjananaidu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

MALWARE

DETECTION
AND
ANALYSIS
(GROUP-20)
Internal Supervisor – Dr. Alekha Kumar Mishra
1.PENTA SANJANA NAIDU(2021UGCS044)
2.SABYASACHI BISWAL(2021UGCS099)
3.ISHAN LAKRA(2021UGCS066)
I. Introduction of problem
II. Literature Survey
Contents III. Research Gap and Motivation
IV. Proposed Possible Solution
V. References
INTRODUCTION
OF PROBLEM
• Malware Detection using Deep Learning Challenge: The
challenge lies in developing deep learning models capable of
effectively identifying diverse and evolving malware threats.

• Rising Malware Threats: Increasing frequency and complexity of

Malware
malware attacks pose a significant cybersecurity challenge.

• Evolution of Malware: Malware variants continuously evolve,


Detection making signature-based detection less effective.

using Deep • Need for Accuracy: Traditional methods often lead to false
positives or fail to detect zero-day threats.
Learning • Scalability: Scalability issues in handling vast malware datasets
and real-time detection.

• Explainability: Lack of interpretability in deep learning models


hinders trust and adoption.

• Resource Intensity: Deep learning models can be resource-


intensive, making them impractical for some applications.
• Critical for Reliability: It is crucial in the field of cybersecurity to safeguard computer systems,
networks, and sensitive data against malicious software threats.
• Essential Solution: Deep Learning technology has become essential to enhance protection
against various malware
• Diverse Approaches: The methods cater to different scenarios, ensuring improved detection of
various malware while taking account of its magnitude.
• Continuous Advancements: Ongoing research and advancements aim to enhance these
techniques, ensuring better performance in various real-world applications.
• Various Deep Learning Methods: Different techniques exist for
detecting malware across a platform to protect and safeguard an
organization or a user.

Literature
• Types of Detection Method: There are various ways of malware
detection using deep learning such as Recurrent Neural

Survey Networks, Long Short-Term Memory, Graph Neural Networks etc.

• Recurrent Neural Networks (RNN): Effective for sequential data,


such as API call sequences in malware behavior analysis.

• Long Short-Term Memory (LSTM): A type of RNN with memory


capabilities, suitable for time-series data in malware analysis.

• Graph Neural Networks (GNN): Suitable for modeling malware


relationships in network structures.
Key Findings
• Deep learning methods consistently outperform traditional approaches.

• Ensemble techniques improve the robustness of malware detection systems.

Conclusion

• Deep learning is a promising avenue for malware detection.


• Ongoing research and innovation are essential to stay ahead of evolving threats.
• Evolving Malware Landscape: The rapid evolution of malware
strains poses a challenge for traditional detection methods, creating a
need for more adaptable solutions.

• Zero-Day Threats: The inability of signature-based systems to detect

Research
zero-day threats calls for advanced techniques like deep learning to
identify previously unknown malware.

Gap or • Real-Time Detection: Timely identification of malware is critical,

Motivation
especially in real-time scenarios. Deep learning can improve the
speed and accuracy of detection.

• Enhanced Security Posture: The overarching motivation is to


bolster cybersecurity efforts by filling the gaps in malware detection
capabilities, making systems more robust and resilient against
evolving threats.
Proposed Idea:
• Introduces a novel approach to malware detection based on Deep
Convolutional Neural Networks (CNNs).
• Utilizes the inherent visual and structural patterns within malware
binaries to enhance detection accuracy.

Proposed Methodology:
• Trains deep CNNs to learn distinctive features from malware binaries.
Possible • Utilizes a comprehensive dataset containing diverse malware
samples.

Solution • Employs transfer learning techniques to improve model


generalization.

Implications:
• Offers a promising avenue for more robust and
adaptable malware detection systems.
• Potential applications in real-time threat detection
and cybersecurity.
REFERENCES

• A Unique Approach to Malware Detection Using Deep Convolutional Neural Networks - Harsh
Malani, Aarush Bhat, Shubham Palriwala, Jyotir Aditya, Anant Chaturvedi (Vellore Institute of
Technology, 2022)
• Malware analysis and multi-label category detection issues: Ensemble-based approaches - Izzat
Alsmadi, Bilal Al-Ahmad, Mohammad Alsmadi (Texas A&M, 2022)
• Detection Of Malware Using Deep Learning - Sabila Newaz, Hasan Md Imran, Xingya Liu (Lamar
University, 2021)
• PRACTICAL MALWARE ANALYSIS - Michael Sikorski and Andrew Honig

You might also like