Adversarial Artificial Intelligence
Developing robust AI systems that withstand adversarial attacks — across network security, mobile malware, healthcare AI, and social network manipulation.
AI systems deployed in healthcare, security, and finance can be manipulated by adversarial inputs — carefully crafted perturbations that cause confident wrong predictions. We develop defenses, detection systems, and robust architectures that remain reliable even when attacked.
Research Areas
- Adversarial CAPTCHAs & Usable Security — Designing CAPTCHAs that are easy for humans but hard for AI using precise noise targeting. Published at AAMAS 2026.
- Encrypted Network Traffic — Detecting anomalies and classifying malicious traffic in fully encrypted streams without decryption, using contrastive learning for zero-day attacks.
- Android Malware Detection — ML-based classifiers hardened against evasion attacks, maintaining accuracy even when adversaries know the detection method.
- Healthcare AI Robustness — Adversarial training and certified defenses for medical diagnosis and patient prediction models under attack.
- Social Network Manipulation — Detecting coordinated inauthentic behavior, bot accounts, and disinformation using graph-based anomaly detection.
Technical Approaches
- Contrastive & Self-Supervised Learning — SimCSE-based methods that build robust representations effective for zero-shot detection of novel attack patterns.
- Adversarial Training & Certified Defenses — Training on adversarial examples with provable robustness guarantees via randomized smoothing for safety-critical applications.
- Privacy-Preserving Detection — Federated learning and differential privacy techniques that protect sensitive data while maintaining detection performance.
Related Publications
2026
- Uncovering Microservice Faults: A Temporal Graph Approach to Root Cause Analysis2026Proceedings of the IEEE International Conference on Communications. ICC 2026
- 2026Computer Networks
- Cleaner Adversarial CAPTCHAs: Intelligent Targets and Precise Noise for Usable Security2026Proceedings of the 25th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2026)
2025
- Cloudy with a Chance of Anomalies: Dynamic Graph Neural Network for Early Detection of Cloud Services’ User Anomalies2025Proceedings of the 34th International Conference on Computer Communications and Networks
- Optimized File Type Detection and One-Shot Reclassification Model2025Proceedings of the IEEE International Conference on Communications
- A New D-MAGIC: Dynamic Model for Cybersecurity Attack Detection Using GNNs into Clustering2025Proceedings of the IEEE International Conference on Communications
- PQClass: Classification of Post-Quantum Encryption Applications in Internet Traffic2025Proceedings of the IEEE International Conference on Communications
- Leveraging OSINT for Advanced Proactive Cybersecurity: Strategies and Solutions2025IEEE Access
2024
- Few-Shot API Attack Detection: Overcoming Data Scarcity with GAN-Inspired Learning2024arXiv preprint arXiv:2405.11258
- Extending Limited Datasets with GAN-Like Self-Supervision for SMS Spam Detection2024Computers & Security
2023
- Breaking the Structure of MaMaDroid2023Expert Systems with Applications
- Detecting Parallel Covert Data Transmission Channels in Video Conferencing Using Machine Learning2023Electronics
2022
- MaMaDroid2.0–The Holes of Control Flow Graphs2022arXiv preprint arXiv:2202.13922
- Problem-Space Evasion Attacks in the Android OS: A Survey2022arXiv preprint arXiv:2205.14576
- Do You Think You Can Hold Me? The Real Challenge of Problem-Space Evasion Attacks2022arXiv preprint arXiv:2205.04293
- Less Is More: Robust and Novel Features for Malicious Domain Detection2022Electronics
- MalDIST: From Encrypted Traffic Classification to Malware Traffic Detection and Classification20222022 IEEE 19th annual consumer communications & networking conference (CCNC)
2021
- Crystal Ball: From Innovative Attacks to Attack Effectiveness Classifier2021IEEE Access
- Robust Coordination in Adversarial Social Networks: From Human Behavior to Agent-Based Modeling2021Network Science
2020
- Encrypted Video Traffic Clustering Demystified2020Computers & Security
- Evasion Is Not Enough: A Case Study of Android Malware2020International symposium on cyber security cryptography and machine learning
- Robust Malicious Domain Detection2020Cyber Security Cryptography and Machine Learning: Fourth International Symposium, CSCML 2020, Be’er Sheva, Israel, July 2–3, 2020, Proceedings 4
2019
- Adversarial Coordination on Social Networks2019Proceedings of the 18th International Conference on Autonomous Agents and Multiagent Systems
- Improving Robustness of ML Classifiers Against Realizable Evasion Attacks Using Conserved Features201928th USENIX Security Symposium (USENIX Security 19)
2018
- Adversarial task assignment2018International Joint Conference on Artificial Intelligence