Kyoungmin Roh Cybersecurity Researcher
Pursuing an academic career in cybersecurity.
Research focus:
- Learning-based IDS under concept drift
- Post-quantum cryptography in CPS and embedded systems
- Secure system design that remains trustworthy over time
Scroll
About
Objective
To pursue an academic career in cybersecurity as a professor, contributing to society through impactful research and education. I aim to develop technologies that enhance security and people's lives while mentoring the next generation of researchers who will advance the field beyond my own contributions.
Research Interests
I am interested in the security of complex systems that must operate reliably under evolving threats and broken assumptions. My current research focuses on CPS and system security, with a particular emphasis on learning-based intrusion detection (IDS) under concept drift and the secure deployment of post-quantum cryptography (PQC) in CPS and embedded systems. My research goal is to establish principled foundations for designing secure systems that remain trustworthy over time, across various environments, and during the post-quantum transition.
Education
Dankook University
B.E. in Cybersecurity
- Undergraduate (Junior → Senior)
- GPA: 3.0 / 4.5
- Relevant coursework: AI Security, CPS Security, Embedded Security, Cryptography
Experience
KATUSA (Korean Augmentation to the U.S. Army)
U.S. 8th Army, Camp Carroll, 35th Brigade, 2-1 Air Defense Artillery Battalion, Echo Company
Publications
SCI Journals
SCAN: Structural Clustering with Adaptive Thresholds for Intelligent and Robust Android Malware Detection under Concept Drift
Computer Modeling in Engineering & Sciences (CMES), 2026 [SCIE Q1] (Accepted)
Domestic Journals
Concept Drift-Resilient Android Malware Detection via API Co-occurrence Graphs and Louvain Communities
KIISE Transactions on Computing Practices (KTCP), 2025 (Invited)
View Paper →International Conferences
ALARM: Android Malware Detection with Leiden API Communities and Robust Mixture of Experts
The 41st ACM/SIGAPP Symposium on Applied Computing (SAC 2026), Thessaloniki, Greece, Mar. 2026 (Accepted)
Domestic Conferences
Drift-Aware Security Module based on Louvain Communities for Retraining-Free Android Malware Detection
Korea Software Congress (KSC 2025), Yeosu, Republic of Korea, Dec. 2025 — Best Paper Award
Android Malware Detection using Co-occurrence Graphs of APIs and Louvain Method for Community Detection
WDSC 2025: Workshop on Dependable and Secure Computing, Jeju, Republic of Korea, Aug. 2025 — Best Paper Award
View Paper →Classifying File Fragment Types for IVI System Forensics
WDSC 2025: Workshop on Dependable and Secure Computing, Jeju, Republic of Korea, Aug. 2025
View Paper →Thesis
Qrust: AI-powered QR Phishing Detection and Secure QR Generation
Bachelor's Degree Thesis, Department of Cybersecurity, Dankook University, Dec. 2025 — Audience Choice Award
View Paper →Other Research Papers
A Lightweight Secret-Isolated Post-Quantum Cryptographic Architecture for ARM TrustZone
Cyber Security Contest, Department of Cybersecurity, Dankook University, Dec. 2025 — 1st Place Award
View Paper →Awards & Honors
Dec 2025
Specialized Project Scholarship
Awarded for strong track record of academic awards and high-quality project execution; recognized for excellence in AI security–oriented research.
Dankook University, Department of Cybersecurity
Dec 2025
Academic Research Scholarship
Published a paper at a leading domestic academic conference as an undergraduate; awarded for research potential and academic excellence.
Dankook University, National Center of Excellence in Software
Dec 2025
Cyber Security Contest 1st Place Award
"A Lightweight Secret-Isolated Post-Quantum Cryptographic Architecture for ARM TrustZone"; isolated secret-based computation in PQC Kyber.
Dankook University, Department of Cybersecurity
Dec 2025
KSC 2025 Best Paper Award
"Drift-Aware Security Module Based on Louvain Communities for Retraining-Free Android Malware Detection"; recognized for robustness and scalability under concept drift.
Korean Institute of Information Scientists and Engineers (KIISE)
Dec 2025
Audience Choice Award, Capstone Festival
Qrust: Secure QR & AI-Based Phishing Detection System; HMAC-signed QR validation, ML-based malicious URL detection.
Dankook University, National Center of Excellence in Software
Sep 2025
Participation Prize (3rd Place), Dankook Startup Hackathon
Lead developer and planner for Ddobak, an LLM-based speech correction app; recognized for innovative AI-driven assistive healthcare design.
Dankook University, National Center of Excellence in Software
Aug 2025
WDSC 2025 Best Paper Award
Android malware detection using API co-occurrence graphs and Louvain communities; selected for robustness under concept drift.
Korean Institute of Information Scientists and Engineers (KIISE)
View certificate →Dec 2024
The Army Commendation Medal (ARCOM)
For commendable service as equipment records and parts clerk KATUSA; supported the unit's mission.
United States Department of the Army
Oct 2024
Best KATUSA Award
Outstanding marksmanship performance; set an example for fellow KATUSA soldiers.
Eighth United States Army Republic of Korea Army Support Group
Sep 2023
1st Place, Turing Cipher Idea Contest
Hybrid post-quantum signature scheme combining Lamport signatures and Merkle trees; evaluated on originality, cryptographic soundness, and feasibility.
Dankook University, Department of Cybersecurity
Jun 2022
Specialized Project Scholarship
Awarded for research initiative in AI security and high-quality project execution.
Dankook University, Department of Cybersecurity
Mar 2021
Admission Scholarship
Awarded for academic excellence and demonstrated leadership potential.
Dankook University
Patent
A Malware Detection Method Combining Clustering and Supervised Learning Models
Projects
Split-Kyber for ARM TrustZone-A
Secret-Isolated PQC · 1st Place, Cyber Security Contest
Implemented a split execution framework for CRYSTALS-Kyber (ML-KEM) on ARM TrustZone-A. Secret-dependent operations run exclusively in Secure World (TEE), ensuring strict key isolation. Aims to minimize the Trusted Computing Base (TCB) while maintaining correctness and performance.
GitHub →LV.0: LLM Vulnerability Zero
LLM-powered Security Vulnerability Reporter
Developed an AI-based static analysis and automated reporting framework for open-source vulnerabilities. Integrated GitHub workflow automation and NLP-based risk summarization. Led backend and AI system design using Flask and FastAPI.
GitHub →ASX: Android API Sequence Extractor
Static Analysis · Electron GUI
Designed a static analysis pipeline to extract API-level call sequences from DEX files. Built preprocessing logic for multi-instance learning (MIL) workflows. Developed as a cross-platform Electron-based GUI tool.
GitHub →Qrust: Secure QR & AI Phishing Detector
Capstone · Audience Choice Award
Developed a secure QR generator and mobile phishing detection system using ML classifiers. Implemented HMAC-signed QR content and Flask-based malicious URL detection API. End-to-end security workflow: scanning → validation → risk scoring.
GitHub →Digital Forensic Toolkit for Hyundai Avante (CN7)
IVI Forensics · Log-based Event Reconstruction
Developed a forensic extraction and log-based event reconstruction system for Android-based IVI systems. Automated vehicle log parsing and visualization for forensic workflows. Designed for researchers analyzing Hyundai vehicle platforms.
GitHub →Post-Quantum Signature System
Lamport + Merkle Tree · 1st Place, Turing Cipher Idea Contest
Implemented a hybrid post-quantum signature system combining Lamport OTS and Merkle tree aggregation. Python-based prototype demonstrating signing, verification, and one-time key management.
GitHub →AI-powered Document Summarization Web Application
FastAPI · Claude API · Docker
Developed a FastAPI-based backend API for user authentication and project management. Implemented PDF/TXT upload and AI summarization via Claude API integration. Provided Swagger UI / ReDoc documentation, SQLite persistence, and Docker-based deployment.
GitHub →Ddobak — LLM-based Speech & Hearing Therapy App
Flutter · 3rd Place, Dankook Startup Hackathon
Developed an LLM-powered pronunciation correction app supporting hearing-impaired users. Led UI/UX design and ideation. Built a Flutter prototype integrated with an AI feedback system.
GitHub →Smart Greenhouse AI
CNN · Transfer Learning · IoT-ready ML
Trained a CNN-based classifier for pest detection in agriculture. Applied image augmentation and transfer learning for robustness. Developed an IoT-ready ML pipeline using Google Colab.
GitHub →Selenium Web Crawler
Automation · Dataset for LLM Research
Built an automation script for scraping product reviews to generate datasets for LLM research. Used Selenium for DOM traversal and text extraction/cleaning.
GitHub →Contact
Phone: +82-10-2506-3409
Location: Seongnam-si, Gyeonggi-do, Republic of Korea