Sotabase

Career

· Assistant Professor, University of California, Berkeley2021–
· Assistant Professor, Cornell University2019–2020
· Postdoctoral Researcher, Microsoft Research2018–2019
· Research Intern, Microsoft Research2016–2016
· PhD in Computer Science, Carnegie Mellon University2013–2018
· Graduate Student, University of Waterloo2011–2013
· Intern, Google2010–2010
· Research Assistant, University of Waterloo2009–2009

Publications (96)

Jailbroken: How Does LLM Safety Training Fail?
Neural Information Processing Systems · 2023
1,483
cited
Commitment Without Regrets: Online Learning in Stackelberg Security Games
ACM Conference on Economics and Computation · 2015
142
cited
Learning Optimal Commitment to Overcome Insecurity
Neural Information Processing Systems · 2014
123
cited
Efficient Learning of Linear Separators under Bounded Noise
Annual Conference Computational Learning Theory · 2015
100
cited
Learning and 1-bit Compressed Sensing under Asymmetric Noise
Annual Conference Computational Learning Theory · 2016
100
cited
Multi-Agent Risks from Advanced AI
arXiv.org · 2025
88
cited
The disparate equilibria of algorithmic decision making when individuals invest rationally
FAT* · 2019
76
cited
Maximizing Welfare with Incentive-Aware Evaluation Mechanisms
International Joint Conference on Artificial Intelligence · 2020
75
cited
Ignorance is Almost Bliss: Near-Optimal Stochastic Matching With Few Queries
ACM Conference on Economics and Computation · 2014
74
cited
Oracle-Efficient Online Learning and Auction Design
IEEE Annual Symposium on Foundations of Computer Science · 2016
67
cited
Collaborative PAC Learning
Neural Information Processing Systems · 2017
66
cited
Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation
International Conference on Machine Learning · 2024
64
cited
Smoothed Analysis with Adaptive Adversaries
IEEE Annual Symposium on Foundations of Computer Science · 2021
61
cited
One for One, or All for All: Equilibria and Optimality of Collaboration in Federated Learning
International Conference on Machine Learning · 2021
59
cited
Smoothed Analysis of Online and Differentially Private Learning
Neural Information Processing Systems · 2020
55
cited
k-center Clustering under Perturbation Resilience
International Colloquium on Automata, Languages and Programming · 2015
51
cited
Online Learning with a Hint
Neural Information Processing Systems · 2017
48
cited
The Provable Virtue of Laziness in Motion Planning
International Conference on Automated Planning and Scheduling · 2017
47
cited
On-Demand Sampling: Learning Optimally from Multiple Distributions
Neural Information Processing Systems · 2022
46
cited
Three Strategies to Success: Learning Adversary Models in Security Games
International Joint Conference on Artificial Intelligence · 2016
42
cited
Sotabase
Nika Haghtalab | Researcher Profile | Sotabase | Sotabase