Cybersecurity : Cyber Defense, Privacy and Cyber Warfare.

Format:

Book

Author/Creator:

Dimitoglou, George.

Series:

Intelligent Computing Series

Intelligent Computing Series ; v.4

Language:

English

Subjects (All):

Computer security.

Computer networks--Security measures.

Computer networks.

Physical Description:

1 online resource (0 pages)

Edition:

1st ed.

Place of Publication:

Berlin/Boston : Walter de Gruyter GmbH, 2025.

Summary:

The depth of our dependence on technology in our interconnected world underscores the critical role of cyber security.This book provides essential insights into the latest research and best practices and techniques for protecting against cyber threats, emphasizing the critical relevance of cyber security in safeguarding personal information.

Contents:

Intro

Preface

Contents

List of contributing authors

Game-based testing for active cyberdefense and cyberdeception

1 Introduction

2 Previous work

2.1 Military cyberspace operations

2.2 Active defenses in cyberspace and cyberdeception

2.3 Game analysis

3 The CCAT tool, an active-defense planner

3.1 Experiment design

3.2 Scenario components and testing

3.3 Possible actions

3.4 Costs and benefits of options

3.5 Training the models

3.6 Average final scores of the CCAT experiments

4 Decepgame, a planner against advanced persistent threats

4.1 Game design

4.1.1 Player generation

4.1.2 Action profiles and player specification generation

4.2 Implementation of the APT game

4.3 Results of the decepgame APT simulation

4.4 Discussion of the decepgame results

5 Precalculating defender tactics from attacker variables

6 Conclusions

References

Graph-ensemble methods for generating malware behavioral signatures

1.1 Natural language processing approaches

1.2 Introduction to graph neural networks

1.3 Application of graph neural networks in anomaly detection

2 Methodology

2.1 Datasets

2.2 Model architectures

2.2.1 Feature concatenation

2.2.2 Embedding concatenation

2.3 Heterogeneous graphs

2.4 Graph feature importance and signatures

2.5 Model performance metrics

3 Results and discussion

3.1 Analysis of malicious behavior

3.2 Analysis of unknown binaries

3.3 Ensemble model complexity

Efficient cyber threat detection on SCADA systems using feature-grouped generative adversarial networks

2 Related works

3 Generative adversarial networks

3.1 Generative models

3.2 GAN structure

3.3 Weaknesses of GANs

3.4 Loss function

4 Measuring performance

5 Evaluation setup.

6 Grouping dataset

7 Model architecture

8 Results

9 Discussion

10 Conclusion

Applying artificial intelligence techniques to intrusion detection systems in serial-based industrial networks

2 Background

2.1 Industrial networks

2.2 Security issues in industrial networks

2.3 Intrusion detection systems (IDSs)

3 AI techniques applied in IDS for industrial networks

4 IDS in industrial serial-based networks

5 Conclusion

A hybrid intelligent intrusion detection system

3 Proposed HIIDS

4 The evaluation and analysis

4.1 Time analysis

4.2 Precision and accuracy

4.3 Data loss

5 Evaluation

6 Conclusion and future work

PwnPilot: could an adversary be pair programming with our most trusted software engineers?

2 Background and related work

2.1 Potential benefits and risks of AI code assistants

2.2 Evaluating correctness, quality, and robustness of AI-generated code

2.3 Evaluating security of AI-generated code

2.4 A surreptitious adversary paired with every programmer?

2.5 Positioning PwnPilot within a taxonomy of generative AI misuse cases

3 PwnPilot: security smells, covert poisoning, and theoretically undetectable backdoors

3.1 Threat analysis: many paths to PwnPilot

4 A new hope: current mitigation options and future research directions

4.1 Building a truly trusted model

4.2 Trusting commercial enhancements

4.3 Automated static code analysis

4.4 AI-driven N-version programming (AID-NVP)

4.5 AI-accelerated automated testing

4.6 Diverse, multi-AI feedback loops

4.7 Improving AI interactions, prompt engineering, and prompt injection prevention

4.8 Using attack techniques for good: model extraction for validation.

4.9 Future symbolic abstractions for robust, secure models

5 Technology readiness level (TRL) of PwnPilot mitigations options in 2024

5.1 An overview of technology readiness levels (TRLs)

5.2 TRLs for PwnPilot mitigation options in 2024

6 Final reflections on trusting trust in the age of AI pair programmers

How to attack a far galaxy and beyond

2.1 Side-channel attacks

2.2 Neural networks

3 CRISTALS-Kyber

3.1 Algorithm

3.2 Determiner leakage attack

3.3 Plaintext checking oracle

3.4 Multi-bit error injection attack

3.5 Other attacks

3.6 AI attacks against Kyber in short

4 CRYSTALS-Dilithium

4.1 Algorithm

4.2 Number-theoretic transform attack

4.3 Bit-unpacking function attack

4.4 Other attacks

4.5 AI attacks against Dilithium in short

5 FALCON

5.1 Algorithm

5.2 Floating point multiplication

5.3 Base sampler

5.4 AI attacks against FALCON in short

6 SPHINCS+

6.1 Algorithm

6.2 One-time signature schemes

6.3 AI attacks against SPHINCS^ {\mplus} in short

7 AI analysis

8 Conclusions

Injecting uniform chaotic sequences into an ANN's learning fabric to reduce overfitting

2.1 Conventional techniques

2.2 Entropy-based techniques

2.3 Chaos injection versus random noise injection

3 Proposed work

3.1 Improved multiparametric tent map (MTM)

3.2 The influence of initial values on uniformity of chaotic sequences

3.3 Generate chaotic values based on dataset parameters

4 Experimental results

5 Conclusions and future work

Effectiveness of machine learning and deep learning in cybersecurity

2 Literature survey

3 Cyber security: application of ML algorithms

3.1 Shallow learning

3.2 Deep learning.

4 Machine learning algorithms applications

5.1 Shallow versus deep learning

5.2 Specific detectors versus general

5.3 Vulnerability to adversarial attacks

5.4 Selection of a machine learning algorithm

6 Conclusion

Quantum-enhanced cyber threat detection with mini-batch optimization

3 Quantum binary classification research in cybersecurity

4 Experimental setup and results

4.1 Experimental setup

4.2 Experimental results

5 Conclusion and future work

Future of auditable AI systems

2 Theoretical basics

2.1 Functionality of neural networks

2.2 Artificial intelligence challenges, attacks, and defenses

2.2.1 Definition

2.2.2 Classification of AI

2.2.3 AI challenges

2.2.4 Defense mechanisms

2.3 Auditing artificial intelligence

2.3.1 Forensic audit

2.3.2 Aims of auditing AI

2.3.3 eXplainable artificial intelligence (XAI)

2.3.4 Current possibilities to audit AI

3 Methods

3.1.1 Implementation of the continual learning model

3.1.2 Logging the weights of a continual learning AI system

3.1.3 Weight extraction

3.1.4 Logging visualization

3.1.5 Hashing the output

4 Results

4.1 Logging the weights of continual learning AI systems

4.1.1 Memory size of the weights log file versus training run

4.1.2 Memory size of the weights log file versus number of neurons

4.1.3 Proportion of zeros in the delta log file versus training run

4.2 Current possibilities of auditing intelligent systems

4.2.1 Memory size of the weights log file versus training run

4.2.2 Memory size of the weights log file versus number of neurons

4.2.3 Proportion of zeros in the delta log file versus training run

5 Discussion

5.1 Current possibilities of auditing artificial intelligence.

5.1.1 eXplainable artificial intelligence

5.1.2 Statistical methods

5.1.3 Digital forensic readiness

5.1.4 Standardization

5.2 Possibilities of logging continual learning

5.2.1 eXplainable artificial intelligence

5.2.2 Auditing autonomous systems

5.2.3 Auditing text and image processing

5.2.4 Summary

5.3 Toward current research

5.3.1 Research projects

5.3.2 Transparency

5.3.3 Forensic audit

5.4 Artificial intelligence act and legal challenges

5.4.1 Risk classification

5.4.2 Logging

6.1 Statistical method

6.2 Forensic audit

6.3 Artificial intelligence act

6.4 Problem statement

Virtual cybersecurity testbeds for industrial Internet of Things

2 Case study

2.1 Method

2.1.1 Virtual industrial testbed for cybersecurity

2.1.2 Denial-of-service (DoS) attack

3 Results

3.1 Virtual industrial testbed for cybersecurity

3.1.1 Testbed setup with VirtualBox

3.1.2 Testbed setup with Hyper-V

3.1.3 Testbed setup with VMware workstation Pro 16

3.2 Denial-of-service (DoS) attack

4 Discussion and conclusion

4.1 Findings

4.2 Future work

4.3 Conclusion

Security verification of authenticated encryption with associated data under chosen message attack assumption using Tamarin prover

1.1 Background

1.2 Contributions

1.2.1 Proposed method

1.2.2 Case studies

1.3 Related works

1.4 Structure of this paper

2 Preliminaries

2.1 Tamarin prover description and verification result

3 Attack models subject to formalization

3.1 Digital signature of MITM attack model

3.2 EUF-CMA model

3.3 IND-CPA model

4 Formal verification of the digital signatures of the MITM attack, EUF-CMA, and IND-CPA models

4.1 Formalization: digital signature of MITM attack model.

4.2 Formalization: EUF-CMA model.

Notes:

Description based on publisher supplied metadata and other sources.

Part of the metadata in this record was created by AI, based on the text of the resource.

ISBN:

3-11-143654-3

9783111436548

OCLC:

1553698953