Sotabase

Career

· Director, Qualcomm2025–
· Professor and Director of the Machine Learning Department, Carnegie Mellon University2024–
· Member of the Board of Directors, OpenAI2024–
· Chief Expert, Bosch2018–
· Associate Professor, Carnegie Mellon University2018–2024
· Chief Data Scientist, C3.ai2014–2018
· Assistant Professor, Carnegie Mellon University2012–2018
· Postdoctoral Researcher, Massachusetts Institute of Technology2010–2012
· PhD in Computer Science, Stanford University2005–2010
· B.S. in Computer Science, Georgetown University2001–2005
· Co-founder and Chief Technical Advisor, Gray Swan AI

Publications (65)

Overfitting in adversarially robust deep learning
International Conference on Machine Learning · 2020
895
cited
Representation Engineering: A Top-Down Approach to AI Transparency
arXiv.org · 2023
739
cited
End-to-End Differentiable Physics for Learning and Control
Neural Information Processing Systems · 2018
438
cited
SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver
International Conference on Machine Learning · 2019
286
cited
Beta-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Neural Network Robustness Verification
Neural Information Processing Systems · 2021
262
cited
Improving Alignment and Robustness with Circuit Breakers
Neural Information Processing Systems · 2024
217
cited
Finetune like you pretrain: Improved finetuning of zero-shot vision models
Computer Vision and Pattern Recognition · 2022
194
cited
On Physical Adversarial Patches for Object Detection
arXiv.org · 2019
193
cited
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
International Conference on Learning Representations · 2024
138
cited
Test-Time Adaptation via Conjugate Pseudo-labels
Neural Information Processing Systems · 2022
127
cited
Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation
European Conference on Computer Vision · 2022
109
cited
Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shift
Neural Information Processing Systems · 2022
107
cited
Sparse Gaussian Conditional Random Fields: Algorithms, Theory, and Application to Energy Forecasting
International Conference on Machine Learning · 2013
105
cited
A Framework for robustness Certification of Smoothed Classifiers using F-Divergences
International Conference on Learning Representations · 2020
73
cited
Contextually Supervised Source Separation with Application to Energy Disaggregation
AAAI Conference on Artificial Intelligence · 2013
73
cited
Efficiently Computing Local Lipschitz Constants of Neural Networks via Bound Propagation
Neural Information Processing Systems · 2022
56
cited
Deep Equilibrium Approaches to Diffusion Models
Neural Information Processing Systems · 2022
53
cited
Not All Rollouts are Useful: Down-Sampling Rollouts in LLM Reinforcement Learning
arXiv.org · 2025
48
cited
Industrial demand response by steel plants with spinning reserve provision
North American Power Symposium · 2015
47
cited
Camera Based Localization for Autonomous UAV Formation Flight
Infotech@Aerospace 2011 · 2011
44
cited
Sotabase
Zico Kolter | Researcher Profile | Sotabase | Sotabase