Bruce Rushing, PhD

Bruce Rushing, PhD, is a Research Scientist within the Data Analytics Center (DAC). The DAC provides machine learning (ML) and artificial intelligence (AI) analytics to UVA researchers as a core part of its mission. Dr. Rushing has expertise in ML and AI, with an emphasis on generative probabilistic modeling, deep learning with neural networks, and large language models (LLMs). His previous work at the DAC includes probabilistic models for glaucoma patient trajectories, next-image segmentation and generation with time-series MRI data, and LLM fine-tuning for modeling semantic shifts in scientific descriptions of American wildlife. Previously, he conducted foundational research on ML, statistics, LLMs, and causal discovery as a Postdoc at Purdue University’s Virtual Reality and Artificial Intelligence lab. He is a trained computational philosopher and statistician, and he has a PhD in Philosophy and MA in Mathematical Behavioral Sciences from the University of California, Irvine, where his dissertation focused on the decision-theoretic foundations for Bayesian and causal inference. His research interests focus on the foundations of statistics and causal modeling and scientific theorizing that emphasize building algorithms for reliable induction. Further information can be found at his website: brucemrushing.com.

Selected Publications
Rushing, B. “Peirce in the Machine: How Mixture of Experts Models Perform Hypothesis Construction,” Philosophy of Science (2025). doi:10.1017/psa.2025.10159
Rushing, B. and Antonios, A. and Espino, H., and Cohen, N., and Baldi, P. “Cold Posterior Effect to Adversarial Robustness,” NeurIPS 2024 Workshop: Bayesian Decision-making and Uncertainty (2024). OpenReview
Rushing, B. “No Free Theory Choice from Machine Learning,” Synthese (2022) doi:10.1007/s11229-022-03901-w

Under Review
Rushing, B., and Gomez-Lavin, J. “Models with a Cause: Causal Discovery with Language Models on Temporally Ordered Text Data”.
Rushing, B., and Danquah, A., and Namezi, A., and Shakeri, H. “A Deep Kernel Learning Approach for Stratifying Glaucoma Trajectories Using Electronic Health Records”.
Rushing, B., and Jain, A., Ruzzetti, E., and Herrmann, D., and Levinstein, B. “Truth Vectors Don’t Steer: A Psychometric Evaluation of Truth Direction Interventions in Language Models”.
Rushing, B. “AI Safety Collides with the Overattribution Bias”.

Education
Ph.D. in Philosophy, University of California, Irvine (July 2023)
M.A. in Mathematical Behavioral Sciences, University of California, Irvine (2022)
M.A. in Philosophy, University of Houston (2017)
B.A. in History, George Washington University (2009)