2024
Limits to scalable evaluation at the frontier: LLM as Judge won't beat twice the data.
CoRR, 2024
Decline Now: A Combinatorial Model for Algorithmic Collective Action.
CoRR, 2024
Lawma: The Power of Specialization for Legal Tasks.
CoRR, 2024
Limits to Predicting Online Speech Using Large Language Models.
CoRR, 2024
Training on the Test Task Confounds Evaluation and Emergence.
CoRR, 2024
ImageNot: A contrast with ImageNet preserves model rankings.
CoRR, 2024
Predictors from causal features do not generalize better to new domains.
CoRR, 2024
Do causal predictors generalize better to new domains?
Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, 2024
An engine not a camera: Measuring performative power of online search.
Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, 2024
Questioning the Survey Responses of Large Language Models.
Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, 2024
Evaluating language models as risk scores.
Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, 2024
Inherent Trade-Offs between Diversity and Stability in Multi-Task Benchmarks.
Proceedings of the Forty-first International Conference on Machine Learning, 2024
Causal Inference from Competing Treatments.
Proceedings of the Forty-first International Conference on Machine Learning, 2024
Allocation Requires Prediction Only if Inequality Is Low.
Proceedings of the Forty-first International Conference on Machine Learning, 2024
Don't Label Twice: Quantity Beats Quality when Comparing Binary Classifiers on a Budget.
Proceedings of the Forty-first International Conference on Machine Learning, 2024
Causal Inference out of Control: Estimating Performativity without Treatment Randomization.
Proceedings of the Forty-first International Conference on Machine Learning, 2024
Test-Time Training on Nearest Neighbors for Large Language Models.
Proceedings of the Twelfth International Conference on Learning Representations, 2024
Unprocessing Seven Years of Algorithmic Fairness.
Proceedings of the Twelfth International Conference on Learning Representations, 2024
2023
Performative Prediction: Past and Future.
CoRR, 2023
What Makes ImageNet Look Unlike LAION.
CoRR, 2023
Difficult Lessons on Social Prediction from Wisconsin Public Schools.
CoRR, 2023
Causal Inference out of Control: Estimating the Steerability of Consumption.
CoRR, 2023
Algorithmic Collective Action in Machine Learning.
Proceedings of the International Conference on Machine Learning, 2023
A Theory of Dynamic Benchmarks.
Proceedings of the Eleventh International Conference on Learning Representations, 2023
Is Your Model Predicting the Past?
Proceedings of the 3rd ACM Conference on Equity and Access in Algorithms, 2023
2022
Algorithmic amplification of politics on Twitter.
Proc. Natl. Acad. Sci. USA, 2022
County-level Algorithmic Audit of Racial Bias in Twitter's Home Timeline.
CoRR, 2022
Backward baselines: Is your model predicting the past?
CoRR, 2022
Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, 2022
Causal Inference Struggles with Agency on Online Platforms.
Proceedings of the FAccT '22: 2022 ACM Conference on Fairness, Accountability, and Transparency, Seoul, Republic of Korea, June 21, 2022
Adversarial Scrutiny of Evidentiary Statistical Software.
Proceedings of the FAccT '22: 2022 ACM Conference on Fairness, Accountability, and Transparency, Seoul, Republic of Korea, June 21, 2022
2021
Patterns, predictions, and actions: A story about machine learning.
CoRR, 2021
Understanding deep learning (still) requires rethinking generalization.
Commun. ACM, 2021
Retiring Adult: New Datasets for Fair Machine Learning.
Proceedings of the Advances in Neural Information Processing Systems 34: Annual Conference on Neural Information Processing Systems 2021, 2021
Alternative Microfoundations for Strategic Classification.
Proceedings of the 38th International Conference on Machine Learning, 2021
From Optimizing Engagement to Measuring Value.
Proceedings of the FAccT '21: 2021 ACM Conference on Fairness, 2021
2020
Revisiting Design Choices in Proximal Policy Optimization.
CoRR, 2020
Stochastic Optimization for Performative Prediction.
Proceedings of the Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, 2020
A System for Massively Parallel Hyperparameter Tuning.
Proceedings of the Third Conference on Machine Learning and Systems, 2020
Test-Time Training with Self-Supervision for Generalization under Distribution Shifts.
Proceedings of the 37th International Conference on Machine Learning, 2020
Balancing Competing Objectives with Noisy Data: Score-Based Classifiers for Welfare-Aware Machine Learning.
Proceedings of the 37th International Conference on Machine Learning, 2020
Proceedings of the 37th International Conference on Machine Learning, 2020
Strategic Classification is Causal Modeling in Disguise.
Proceedings of the 37th International Conference on Machine Learning, 2020
Identity Crisis: Memorization and Generalization Under Extreme Overparameterization.
Proceedings of the 8th International Conference on Learning Representations, 2020
Explaining an increase in predicted risk for clinical alerts.
Proceedings of the ACM CHIL '20: ACM Conference on Health, 2020
Linear Dynamics: Clustering without identification.
Proceedings of the 23rd International Conference on Artificial Intelligence and Statistics, 2020
Generalization in Overparameterized Models.
Proceedings of the Beyond the Worst-Case Analysis of Algorithms, 2020
2019
Strategic Adaptation to Classifiers: A Causal Perspective.
CoRR, 2019
Test-Time Training for Out-of-Distribution Generalization.
CoRR, 2019
Identity Crisis: Memorization and Generalization under Extreme Overparameterization.
CoRR, 2019
A Meta-Analysis of Overfitting in Machine Learning.
Proceedings of the Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, 2019
Model Similarity Mitigates Test Set Overuse.
Proceedings of the Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, 2019
Natural Analysts in Adaptive Data Analysis.
Proceedings of the 36th International Conference on Machine Learning, 2019
The Implicit Fairness Criterion of Unconstrained Learning.
Proceedings of the 36th International Conference on Machine Learning, 2019
The advantages of multiple classes for reducing overfitting from test set reuse.
Proceedings of the 36th International Conference on Machine Learning, 2019
Proceedings of the 7th International Conference on Learning Representations, 2019
Model Reconstruction from Model Explanations.
Proceedings of the Conference on Fairness, Accountability, and Transparency, 2019
The Social Cost of Strategic Classification.
Proceedings of the Conference on Fairness, Accountability, and Transparency, 2019
Open Problem: How fast can a multiclass test set be overfit?
Proceedings of the Conference on Learning Theory, 2019
2018
Gradient Descent Learns Linear Dynamical Systems.
J. Mach. Learn. Res., 2018
Massively Parallel Hyperparameter Tuning.
CoRR, 2018
Group calibration is a byproduct of unconstrained learning.
CoRR, 2018
When Recurrent Models Don't Need To Be Recurrent.
CoRR, 2018
Sanity Checks for Saliency Maps.
Proceedings of the Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems 2018, 2018
Delayed Impact of Fair Machine Learning.
Proceedings of the 35th International Conference on Machine Learning, 2018
2017
Climbing a shaky ladder: Better adaptive risk estimation.
CoRR, 2017
Avoiding Discrimination through Causal Reasoning.
Proceedings of the Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, 2017
Understanding deep learning requires rethinking generalization.
Proceedings of the 5th International Conference on Learning Representations, 2017
Identity Matters in Deep Learning.
Proceedings of the 5th International Conference on Learning Representations, 2017
2016
Private Spectral Analysis.
Encyclopedia of Algorithms, 2016
Special Section on the Fifty-Fourth Annual IEEE Symposium on Foundations of Computer Science (FOCS 2013).
SIAM J. Comput., 2016
Equality of Opportunity in Supervised Learning.
Proceedings of the Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems 2016, 2016
Strategic Classification.
Proceedings of the 2016 ACM Conference on Innovations in Theoretical Computer Science, 2016
Train faster, generalize better: Stability of stochastic gradient descent.
Proceedings of the 33nd International Conference on Machine Learning, 2016
2015
Tight Bounds for Learning a Mixture of Two Gaussians.
Proceedings of the Forty-Seventh Annual ACM on Symposium on Theory of Computing, 2015
Preserving Statistical Validity in Adaptive Data Analysis.
Proceedings of the Forty-Seventh Annual ACM on Symposium on Theory of Computing, 2015
Generalization in Adaptive Data Analysis and Holdout Reuse.
Proceedings of the Advances in Neural Information Processing Systems 28: Annual Conference on Neural Information Processing Systems 2015, 2015
Differentially Private Learning of Structured Discrete Distributions.
Proceedings of the Advances in Neural Information Processing Systems 28: Annual Conference on Neural Information Processing Systems 2015, 2015
The Ladder: A Reliable Leaderboard for Machine Learning Competitions.
Proceedings of the 32nd International Conference on Machine Learning, 2015
2014
Sharp bounds for learning a mixture of two gaussians.
CoRR, 2014
The Noisy Power Method: A Meta Algorithm with Applications.
Proceedings of the Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Processing Systems 2014, 2014
Preventing False Discovery in Interactive Data Analysis Is Hard.
Proceedings of the 55th IEEE Annual Symposium on Foundations of Computer Science, 2014
Understanding Alternating Minimization for Matrix Completion.
Proceedings of the 55th IEEE Annual Symposium on Foundations of Computer Science, 2014
Fast matrix completion without the condition number.
Proceedings of The 27th Conference on Learning Theory, 2014
Computational Limits for Matrix Completion.
Proceedings of The 27th Conference on Learning Theory, 2014
2013
Privately Releasing Conjunctions and the Statistical Query Barrier.
SIAM J. Comput., 2013
On the Provable Convergence of Alternating Minimization for Matrix Completion.
CoRR, 2013
How robust are linear sketches to adaptive inputs?
Proceedings of the Symposium on Theory of Computing Conference, 2013
Beyond worst-case analysis in private singular vector computation.
Proceedings of the Symposium on Theory of Computing Conference, 2013
Multiple Kernel Completion and its application to cardiac disease discrimination.
Proceedings of the 10th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 2013
Algorithms and Hardness for Robust Subspace Recovery.
Proceedings of the COLT 2013, 2013
Robust subspace iteration and privacy-preserving spectral analysis.
Proceedings of the 51st Annual Allerton Conference on Communication, 2013
2012
Can We Reconcile Robustness and Efficiency in Unsupervised Learning?
CoRR, 2012
Beating randomized response on incoherent matrices.
Proceedings of the 44th Symposium on Theory of Computing Conference, 2012
Private data release via learning thresholds.
Proceedings of the Twenty-Third Annual ACM-SIAM Symposium on Discrete Algorithms, 2012
A Simple and Practical Algorithm for Differentially Private Data Release.
Proceedings of the Advances in Neural Information Processing Systems 25: 26th Annual Conference on Neural Information Processing Systems 2012. Proceedings of a meeting held December 3-6, 2012
Proceedings of the Innovations in Theoretical Computer Science 2012, 2012
Fairness through awareness.
Proceedings of the Innovations in Theoretical Computer Science 2012, 2012
2011
A Study of Privacy and Fairness in Sensitive Data Analysis
PhD thesis, 2011
Subsampling Mathematical Relaxations and Average-case Complexity.
Proceedings of the Twenty-Second Annual ACM-SIAM Symposium on Discrete Algorithms, 2011
2010
On the geometry of differential privacy.
Proceedings of the 42nd ACM Symposium on Theory of Computing, 2010
A Multiplicative Weights Mechanism for Privacy-Preserving Data Analysis.
Proceedings of the 51th Annual IEEE Symposium on Foundations of Computer Science, 2010
2009
Deterministically testing sparse polynomial identities of unbounded degree.
Inf. Process. Lett., 2009
Subsampling Semidefinite Programs and Max-Cut on the Sphere.
Electron. Colloquium Comput. Complex., 2009
The uniform hardcore lemma via approximate Bregman projections.
Proceedings of the Twentieth Annual ACM-SIAM Symposium on Discrete Algorithms, 2009
2008
Asymptotically Optimal Hitting Sets Against Polynomials.
Proceedings of the Automata, Languages and Programming, 35th International Colloquium, 2008
Rounding Parallel Repetitions of Unique Games.
Proceedings of the 49th Annual IEEE Symposium on Foundations of Computer Science, 2008
2006
Higher-Order Syntax and Saturation Algorithms for Hybrid Logic.
Proceedings of the International Workshop on Hybrid Logic, 2006