AI × Cybersecurity

Insup Lee

Ph.D. Candidate in Cybersecurity

Korea University  ·  Seoul & Abu Dhabi

Designing domain-specific generative models (DiT, GAN) to address data insufficiency issues in cybersecurity, while investigating vulnerabilities in LLMs & agentic AI.

Insup Lee

About Me

I am an AI & Security Researcher based in both the UAE and South Korea. My approach to security research is built on five years of experience at the Agency for Defense Development (ADD), as well as international joint projects with the UAE Ministry of Defense.

As a Ph.D. candidate in Cybersecurity at Korea University, I have mainly explored how generative models can improve cybersecurity, designing diffusion transformers (DiT) and GANs to address data insufficiency in physical and wireless security, including RF signal processing and side-channel analysis.

For my next steps, I plan to investigate the flip side: uncovering the inherent vulnerabilities within LLMs and agentic AI systems, applying my structural understanding of generative models to ensure their trustworthiness.

Reviewer:
IEEE TDSC IEEE TIFS IEEE TCOM IEEE JSAC

Research

My research lies at the intersection of AI and cybersecurity.

Generative Models for Security

Designing security-specific generative models (diffusion transformers, GANs) to address data insufficiency in physical and wireless security, including RF signal processing and side-channel analysis (SCA).

Diffusion Transformers GANs Side-Channel Analysis RF Security

Cyber Threat Intelligence

Advancing CTI using NLP and LLMs to analyze attack campaigns: TTP extraction, attribution, and generating campaign variants for group identification.

NLP / LLM MITRE ATT&CK TTP Extraction APT Attribution

Securing AI

Transitioning toward securing AI itself, applying structural understanding of generative models to investigate vulnerabilities within LLMs and agentic AI systems, and to ensure their trustworthiness.

LLM Security Agentic AI Adversarial ML AI Safety

Education

Sep 2019 – Present

Ph.D. Candidate in Cybersecurity

Korea University, Seoul, Republic of Korea

Domain-Specific Generative Models for Data Augmentation in Multi-Layer Cybersecurity
Advisors: Prof. Sangjin Lee & Prof. Seokhie Hong  |  Expected: Aug 2026

Mar 2014 – Feb 2018

B.E. in Cyber Defense

Korea University, Seoul, Republic of Korea

Full Tuition Scholarship, Ministry of National Defense  |  Advisor: Prof. Jongin Lim

Employment History

Sep 2025 – Feb 2026

Lecturer

Korea University, Seoul, Republic of Korea

Taught graduate-level course in Computer Networks.

Mar 2025 – Jun 2025

Research Intern

Indiana University Bloomington, IN, USA (Remote)

Explored adversarial attacks on ML systems in autonomous vehicles (Advisor: Prof. Hyungsub Kim).

Aug 2023 – May 2025

Research Engineer (Cyber Officer)

Ministry of National Defense, Republic of Korea / UAE

  • Led international joint research on AI-based security with the UAE Ministry of Defense as PI
  • Investigated AI-based RF signal analysis for drone security [J6]
Jul 2018 – Jul 2023

Researcher

Agency for Defense Development (ADD), Seoul, Republic of Korea

  • Conducted AI-based security research and in-house software development (PI/Mentor: Prof. Changhee Choi)
  • Analyzed nation-sponsored cyber attacks using NLP technologies [J1–J3, J5]
  • Applied generative models for cybersecurity data augmentation [C1, C2]

Publications

C3 2026
DAC Multicore

Exploiting Per-Core Leakage: Electromagnetic Side-Channel Monitoring of Multicore Architectures

D. Bae, S. Park, I. Lee, Y. Jung, K. Lee, H. Kim, and S. Hong

ACM/IEEE Design Automation Conference (DAC), Jul. 2026

BK21+ IF: 3  ·  Top-tier EDA/Hardware Conference

EM Side-ChannelMulticoreHardware Security
J8 2026
LeakDiT

LeakDiT: Diffusion Transformers for Trace-Augmented Side-Channel Analysis

I. Lee, D. Bae, S. Hong, and S. Lee

IEEE Computer Architecture Letters, Vol. 25, No. 1, pp. 5–8, Jan./Jun. 2026

Diffusion Transformer Side-Channel Analysis Data Augmentation
J7 2025
LLM TTP

Multi-Step LLM Pipeline for Enhancing TTP Extraction in Cyber Threat Intelligence

H. Kim, D. Lee, I. Lee, S. Lee, and S. Lee

IEEE Access, Vol. 13, pp. 179696–179710, Oct. 2025

LLMCTITTP Extraction
J6 2025
Drone SPL

Enhancing Modulation Classification via Diffusion Transformers for Drone Video Signal Processing

I. Lee, K. Alteneiji, and M. Alghfeli

IEEE Signal Processing Letters, Vol. 32, pp. 3325–3329, Aug. 2025

Diffusion TransformerDrone SecurityRF Classification
J5 2025
MuCamp

MuCamp: Generating Cyber Campaign Variants via TTP Synonym Replacement for Group Attribution

I. Lee and C. Choi

IEEE Transactions on Information Forensics and Security (TIFS), Vol. 20, pp. 6162–6174, Jun. 2025

JCR 2024 Top 7.8%  in CS, Theory & Methods  ·  Top-tier Security Journal

CTITTP AugmentationAPT Attribution
J4 2024
UniQGAN

UniQGAN: Towards Improved Modulation Classification With Adversarial Robustness Using Scalable Generator Design

I. Lee and W. Lee

IEEE Transactions on Dependable and Secure Computing (TDSC), Vol. 21, No. 2, pp. 732–745, Mar./Apr. 2024

JCR 2023 Top 4.9%  in CS, Software Engineering  ·  Top-tier Security Journal

GANAdversarial RobustnessModulation Classification
J3 2023
Camp2Vec

Camp2Vec: Embedding Cyber Campaign With ATT&CK Framework for Attack Group Analysis

I. Lee and C. Choi

ICT Express, Vol. 9, No. 6, pp. 1065–1070, Dec. 2023

MITRE ATT&CKCampaign Embedding
J2 2023
GloVe TTP

Exploiting TTP Co-occurrence via GloVe-Based Embedding With ATT&CK Framework

C. Shin, I. Lee, and C. Choi

IEEE Access, Vol. 11, pp. 100823–100831, Sep. 2023

GloVeTTP Embedding
J1 2023
BAN

BAN: Predicting APT Attack Based on Bayesian Network With MITRE ATT&CK Framework

Y. Kim, I. Lee, H. Kwon, G. Lee, and J. Yoon

IEEE Access, Vol. 11, pp. 91949–94968, Aug. 2023

Bayesian NetworkAPT Prediction
C2 2019
ICMLA

Anomaly Dataset Augmentation Using Sequence Generative Models

S. Shin, I. Lee, and C. Choi

IEEE International Conference on Machine Learning and Applications (ICMLA), Dec. 2019

SeqGANData Augmentation
C1 2019
ICTC

Opcode Sequence Amplifier Using Sequence Generative Adversarial Networks

C. Choi, S. Shin, and I. Lee

International Conference on ICT Convergence (ICTC), Oct. 2019

SeqGANMalware Detection
Review 2026

Multi-Domain Side-Channel Analysis for Anomaly Detection in Embedded System

S. Park, D. Bae, I. Lee, J. Kim, H. Oh, H. Kim, and S. Hong

Submitted to IEEE Embedded Systems Letters

Side-ChannelAnomaly Detection
Review 2026

(Blind Review)

J. Baek, G. Ahn, S. Park, D. Bae, G. Kim, I. Lee, H. Kim, and S. Hong

Submitted to ACM CCS 2026

ACM CCS
Patent 2024

Method for Augmenting Cyber Attack Campaign Data to Identify Attack Group, and Security

C. Choi and I. Lee

Korea Patent Application No. 10-2024-0176082, Dec. 2024

Patent 2024

Information Identification Method and Electronic Apparatus Thereof

C. Choi, I. Lee, C. Shin, and S. Lee

Korea Patent Application No. 10-2024-0006106, Jan. 2024

Patent 2023

Method for Training Attack Prediction Model and Device Therefor

C. Choi, C. Shin, S. Shin, S. Seo, and I. Lee

U.S. Patent No. US20230308462A1, Sep. 2023

Patent 2021

Apparatus, Method, Computer-readable Storage Medium and Computer Program for Generating Operation Code

C. Choi, S. Shin, and I. Lee

Korea Patent No. 10-2246797, Apr. 2021

Conf. 2025

EM-Based Anomaly Detection using a Dual-Domain Approach

S. Park, D. Bae, I. Lee, H. Kim, and S. Hong

KIISC Winter Conference (CISC-W'25), Nov. 2025

Outstanding Paper Award

Conf. 2025

A Statistical Time-Domain Approach to Anomaly Detection for Robotic-Arm MCU

S. Park, D. Bae, I. Lee, H. Kim, and S. Hong

KIMST Fall Conference, Nov. 2025

Conf. 2025

Reinforcement Learning for Parameter Optimization in CADO-NFS Polynomial Selection

J. Kim, I. Lee, C. Jeon, S. Kim, S. Hong, and S. Lee

KIISC Winter Conference (CISC-W'25), Nov. 2025

Journal 2025

Enhanced DDoS Detection via Traffic Volume-Based Labeling and Transfer Learning

H. Park and I. Lee

Journal of Internet Computing and Services (JICS), Vol. 26, No. 4, pp. 1–8, Aug. 2025

Journal 2024

User Behavior Embedding via TF-IDF-BVC for Web Shell Detection

K. Kim and I. Lee

Journal of The Korea Institute of Information Security & Cryptology (JKIISC), Vol. 34, No. 6, pp. 1231–1238, Dec. 2024

Awards & Honors

Feb 2026

KU Graduate School Achievement Award

Korea University, Seoul, Republic of Korea

View Certificate
Nov 2025

Outstanding Paper Award

CISC-W'25, KIISC

Mar 2025

Certificate of Commendation

UAE Ministry of Defense (UAE-ROK Engagement Program)

View Certificate
Mar 2025

Ambassador's Commendation

Embassy of the Republic of Korea to the UAE

View Certificate
2014 – 2018

Full Tuition Scholarship

Ministry of National Defense, Republic of Korea

Technical Skills

AI & Deep Learning

Diffusion Transformers (DiT) GANs LLM Pipelines Adversarial Robustness NLP

Cybersecurity

CTI / TTP Extraction Side-Channel Analysis RF Signal Analysis Cryptographic Engineering Penetration Testing

Languages & Tools

Python C / C++ CUDA PyTorch Linux Git Docker Streamlit

Contact

I am open to academic collaborations, postdoctoral opportunities, and research discussions in AI security and cybersecurity. Feel free to reach out!

Download Full CV