Uncertainty Representations

Jessica Hullman joins us to share her expertise on data visualization and communication of data in the media. We discuss Jessica’s work on visualizing uncertainty, interviewing visualization designers on why they don't visualize uncertainty, and modeling interactions with visualizations as Bayesian updates.

Guest

Jessica Hullman: I am currently Ginni Rometty Professor of Computer Science and Faculty Fellow at the Institute for Policy Research at Northwestern University. My research develops methods for evaluating and improving AI and ML systems for human moderated decision-making and scientific inference. My students and I study when predictions, explanations, uncertainty estimates, model-generated judgments, or other AI outputs provide valid and useful evidence for downstream decisions. Current work spans AI-assisted decision-making, human-AI and multi-agent complementarity, LLMs as behavioral evidence, and AI systems for scientific research and data analysis. I work between theory and application, grounding my contributions in formal models of rational inference like Bayesian decision theory while addressing applied problems in AI evaluation and deployment. I also maintain an active interest in metascience and statistical reform. I direct the Epistemic Decisions Lab at Northwestern, where I work with many great students and collaborators. My work has been awarded with multiple best paper and honorable mention awards at top conferences. I was chosen as a Microsoft Faculty Fellow (2019), and have been funded by NSF CAREER, Medium, and Small awards, among others. I frequently speak and blog on topics related to uncertainty quantification, AI/ML decision making, statistical modeling, and human-computer interaction.