Collaborate
My research runs across three active lines in trustworthy AI: formal verification, adversarial machine learning, and online safety. Each is an ongoing project with a lab, a PI, and open problems worth working on. If any of the directions below overlap with your interests, I am glad to hear from you.
Online Safety & Social Computing (BullyBlocker Lab)
NSF Secure and Trustworthy Cyberspace (SaTC) & industry research award · PI: Dr. Yasin N. Silva
Active work here focuses on detecting cyberbullying and coordinated harmful behavior in online conversations using full conversational context rather than isolated messages. Current and upcoming projects extend this toward LLM-based systems for conflict detection and mediation at platform scale. Relevant backgrounds include NLP, social computing, and applied machine learning for content moderation.
Contact the PI: ysilva1@luc.edu and CC me at marslan@cs.luc.edu
Adversarial ML & Formal Methods (AISeC / AI4FM Labs)
Two directions run out of this lab. In adversarial machine learning, open problems involve attack and defense in federated learning systems used in healthcare, finance, and infrastructure, as well as measuring the robustness of large language models under adversarial inputs. In formal methods, the focus is on AI-assisted specification writing and automated repair of TLA+ specifications, with the goal of making formal verification tractable for teams without dedicated verification expertise.
Contact the PI: mabuhamad@luc.edu and CC me at marslan@cs.luc.edu
Direct Collaboration
If you have a research direction that sits at the edge of verification, security, or online safety and want to think through it together, write to me directly at marslan@cs.luc.edu