Sotabase

Career

· Research Scientist, California Institute of Technology (Caltech) Center for the Mathematics of Information (CMI)2024–
· Research Lead, AI Safety & Security, ServiceNow Research2024–2025
· PhD Student, University of Washington2024–
· Researcher, Google DeepMind2017–2024
· Researcher, Pacific Northwest National Laboratory2016–2017
· Research Assistant, NEC Laboratories America2012–2012
· Research Assistant, Los Alamos National Laboratory2011–2014
· Postdoctoral Researcher, California Institute of Technology

Publications (139)

On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Models
arXiv.org · 2018
603
cited
Safe Exploration in Continuous Action Spaces
arXiv.org · 2018
484
cited
A Dual Approach to Scalable Verification of Deep Networks
Conference on Uncertainty in Artificial Intelligence · 2018
408
cited
Adversarial Robustness through Local Linearization
Neural Information Processing Systems · 2019
325
cited
A Fine-Grained Analysis on Distribution Shift
International Conference on Learning Representations · 2021
246
cited
Real-Time Optimal Power Flow
IEEE Transactions on Smart Grid · 2017
229
cited
Achieving Verified Robustness to Symbol Substitutions via Interval Bound Propagation
Conference on Empirical Methods in Natural Language Processing · 2019
174
cited
(Certified!!) Adversarial Robustness for Free!
International Conference on Learning Representations · 2022
173
cited
Training verified learners with learned verifiers
arXiv.org · 2018
172
cited
Inverse Optimal Control with Linearly-Solvable MDPs
International Conference on Machine Learning · 2010
165
cited
Stealing Part of a Production Language Model
International Conference on Machine Learning · 2024
141
cited
Rich Human Feedback for Text-to-Image Generation
Computer Vision and Pattern Recognition · 2023
132
cited
Learning Optimal Conformal Classifiers
International Conference on Learning Representations · 2021
113
cited
Enhancing the reliability and accuracy of AI-enabled diagnosis via complementarity-driven deferral to clinicians
Nature Medicine · 2023
102
cited
Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming
Neural Information Processing Systems · 2020
101
cited
VALUATION : A N ADVERSARIAL APPROACH TO UNCOVER CATASTROPHIC FAILURES
2018
86
cited
Interactive Concept Bottleneck Models
AAAI Conference on Artificial Intelligence · 2022
76
cited
A Framework for robustness Certification of Smoothed Classifiers using F-Divergences
International Conference on Learning Representations · 2020
73
cited
Constructing Convex Inner Approximations of Steady-State Security Regions
IEEE Transactions on Power Systems · 2017
64
cited
The Autoencoding Variational Autoencoder
Neural Information Processing Systems · 2020
64
cited
Sotabase
Dvijotham Krishnamurthy | Researcher Profile | Sotabase | Sotabase