AI CyberSafe Innovations Lab
AI CyberSafe Innovations Lab (AICSIL) is an academic research lab dedicated to advancing AI-assisted software security and proactive cyber defense. Our research develops and evaluates AI-driven methods for identifying and assessing software-security risks and proactive defense techniques that make software systems and networks more resilient to cyber threats. We work at the intersection of artificial intelligence, machine learning, optimization, and applied cybersecurity to develop practical and rigorous approaches to emerging security challenges.
Recent updates
Residual Risk Assessment accepted at IEEE SecAssure 2026
Our paper on a unified framework called Residual Risk Scoring quantifies residual risks in patched functions, with PhD student Farhad, was accepted at IEEE SecAssure 2026, co-located with ESORICS.
HYDRA accepted at IEEE AST 2026
Our paper on predicting latent zero-day vulnerabilities in patched functions, with PhD students Farhad and Sabbir, was accepted at IEEE Automated Software Testing 2026.
Malware detection paper accepted at CNC
A comparative evaluation of TabPFN and ensemble models for memory-based malware detection under limited data, with Valentin Leroy and Sharif Ullah, accepted at the 2026 Workshop on Computing, Networking and Communications.
Invited speaker, ICAI 2025
Dr. Dass presented at the International Conference on AI.
Bio-inspired RL paper published in JAICC
With Niloofar Heidarikohol: "Advancing Moving Target Strategy with Bio-Inspired Reinforcement Learning to Secure Misconfigured Software Applications."
ZeroAware
AI-integrated zero-day vulnerability prediction in CI/CD pipelines for critical infrastructure.
Certified Ethical Hacker
Dr. Dass holds the CEH credential and applies it directly in the lab's offensive-security coursework.
Ethical Hacking Club
Dr. Dass founded and advises UL Lafayette's student Ethical Hacking Club.
Making software systems secure and trustworthy
Configuration Hardening
An AI-powered automated configuration management system that combines efficient hardening with proactive defense against attacks on misconfigured applications.
LLM Security in Finance
Examining how large language models can be adopted securely and ethically across banking, where automation meets sensitive financial decision-making.
LLM Security in Medical Applications
Studying privacy and security risks introduced when LLMs are integrated into Electronic Health Records and clinical workflows.
Biometric Authentication Security
Building robust facial-recognition verification for online banking while defending against adversarial machine learning attacks on biometric systems.
Looking for 1–2 PhD / postdoc students
We're recruiting motivated researchers with a background in AI, cybersecurity, or data science to join AICSIL starting Spring/Fall 2025.