I study how humans and machine learning systems can learn from each other through interactive feedback and explainable systems that support human decision-making.
My work sits at the intersection of interactive machine learning and human-factors research: how people and AI systems can hand information back and forth — as feedback, as explanation, and as structured knowledge — without losing trust or accuracy along the way.
Understanding how to effectively integrate human feedback into machine learning workflows, and how the framing of that feedback influences user trust and behavior, from human-subjects evaluations to developing novel explanations of feedback usage.
Why users trust or distrust AI systems, including how first impressions anchor perceptions of accuracy, and how explanations from probabilistic models affect performance on real tasks — largely conducted under the DARPA XAI program.
Tools that turn unstructured text and analyst sessions into structured, explorable knowledge through NLP, feedback-driven knowledge graph recommenders, and summarization pipelines built with intelligence analysts as end users.