About me
Steven Sun is a Ph.D. candidate in Information Technology at George Mason University, advised by Dr. Ray Hong. His work focuses on human-computer interaction, human-AI collaboration, and machine learning.
Research Statement
Research Focus
Steven's research addresses a critical challenge in AI deployment: unexpected model failures can carry severe consequences for safety, productivity, and public trust in AI systems. Despite the urgency of understanding these vulnerabilities, diagnosing and resolving them remains demanding work — requiring significant time, specialized expertise, and access to large-scale data — burdens that fall disproportionately on machine learning engineers.
Approach
To address this, Steven designs novel human-AI collaboration systems that help ML engineers investigate and remediate AI vulnerabilities more efficiently. His work is grounded in two methodological pillars: Scalable Human-In-The-Loop interaction, which sustains meaningful human oversight even under high task loads, and Actionable Interactive Model-Steering, which enables practitioners to translate insights directly into model corrections through intuitive interfaces. His research draws on methods from interactive machine learning, visual analytics, and user-centered design.
Impact
Steven's work contributes to more reliable, trustworthy, and controllable AI across diverse data modalities — including images, time series, and text. By empowering practitioners with better tools for model auditing and correction, he aims to reduce the real-world harm of AI failures and strengthen the human role in the machine learning lifecycle.