Research and publications
My published research examines advanced persistent threats, IoT botnet defenses, and phishing prevention. I am also interested in evaluating AI systems and how AI can support cybersecurity decisions and security testing. I want to understand what these systems do well, where they fail, and what evidence we need before relying on their recommendations.
Published research
My published work covers cyber threat intelligence, IoT botnet defenses, and phishing prevention.
Bryon Miller
A cyber threat intelligence study of twelve APT groups attributed to China, Iran, North Korea, and Russia, followed by six defensive priorities.
Best Paper Award, Research Category, IACIS 2020
Bryon Miller and Xihui Zhang
A review of 21 studies that organizes IoT botnet defenses across ISP architecture, technical detection and mitigation, and user behavior.
Bryon Miller, Katelin Miller, Xihui Zhang, and Mark G. Terwilliger
A literature-based prevention strategy combining one-time passwords, multi-level desktop barriers, and behavior change.
Research projects
Prototype development paused
An AI evaluation prototype for studying mediation in simulated crises. The project received an award in ChinaTalk’s 2026 Evals for the Situation Room contest. Development is currently paused.
The project includes a runnable prototype, engineering checks, and exploratory runs. Those runs do not establish a mediation benefit; scenario review and further controlled testing are needed before drawing stronger conclusions.
The project site presents the earlier design and an illustrative walkthrough.
Questions in scope
This program grows out of my work in cybersecurity research, curriculum, and applied technical education.
How do models respond to incomplete information, conflicting goals, or misleading claims? The Third Chair was developed to study these questions through simulated crises and mediation.
Where can AI help identify weak assumptions or overlooked attack paths in conventional systems? That includes testing the AI’s recommendations, not simply accepting them because they sound reasonable.
What would justify relying on an AI-supported security decision? Accuracy matters, but so do the evidence behind a recommendation, the uncertainty that remains, and the consequences of getting it wrong.
Dissertation research · In progress
My dissertation research examines how a locally run AI system could support defensive decisions in an isolated cyber range. The workflow connects observations from simulated attacks to MITRE ATT&CK techniques and uses MITRE D3FEND to identify possible defensive responses, with a human reviewing recommendations and retaining control of actions.
The evaluation asks whether those recommendations can help interrupt an attack’s progress or prevent it from achieving its objective. This is work in progress, not a published finding or a claim that the approach has been proven effective.
Additional research work
I also work as an on-call Research Engineer with the University of Alabama in Huntsville’s Center for Cybersecurity Research and Education (CCRE). My work includes creating and reviewing cybersecurity curriculum, labs, assessments, and instructional support materials.