Full Professor, University of Minnesota, Minneapolis
3 papers at NeurIPS 2025
We characterize key factors for good synthetic priors and build a SOTA tabular foundation model by pretraining on these priors.
MLZero automates end-to-end multimodal machine learning with high success rate and solution quality without human intervention.
We identify critical flaws in existing datasets and benchmarking protocols for crystal structure prediction in generative modeling of inorganic crystals; we revise datasets and make new benchmarks that account for crystal polymorphism.