Gautam Kamath's articles on arXiv
See also ORCID
https://orcid.org/0000-0003-0048-2559.
- [1] arXiv:2405.20769 [pdf, other]
-
Title: Avoiding Pitfalls for Privacy Accounting of Subsampled Mechanisms under CompositionSubjects: Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [2] arXiv:2406.17216 [pdf, other]
-
Title: Machine Unlearning Fails to Remove Data Poisoning AttacksComments: Published at ICLR 2025Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computers and Society (cs.CY)
- [3] arXiv:2402.00267 [pdf, ps, other]
-
Title: Not All Learnable Distribution Classes are Privately LearnableComments: Appeared in ALT 2024. Added clarification about result, and updated affiliation and funding for VSSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [4] arXiv:2409.19798 [pdf, other]
-
Title: Membership Inference Attacks Cannot Prove that a Model Was Trained On Your DataComments: position paper at IEEE SaTML 2025Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
- [5] arXiv:2503.01766 [pdf, ps, other]
-
Title: Optimal Differentially Private Sampling of Unbounded GaussiansComments: 47 pagesSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [6] arXiv:2412.21061 [pdf, other]
-
Title: BridgePure: Revealing the Fragility of Black-box Data ProtectionComments: 26 pages,13 figuresSubjects: Machine Learning (cs.LG)
- [7] arXiv:2412.02670 [pdf, ps, other]
-
Title: The Broader Landscape of Robustness in Algorithmic StatisticsAuthors: Gautam KamathSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Statistics Theory (math.ST)
- [8] arXiv:2004.00010 [pdf, other]
-
Title: The Discrete Gaussian for Differential PrivacyComments: Correcting a mistake in the statement of Fact 18: this only applies for {\mu} is a half-integer. (This does not affect the results in the paper, which all used {\mu}=0.)Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [9] arXiv:2301.13334 [pdf, other]
-
Title: A Bias-Accuracy-Privacy Trilemma for Statistical EstimationAuthors: Gautam Kamath, Argyris Mouzakis, Matthew Regehr, Vikrant Singhal, Thomas Steinke, Jonathan UllmanSubjects: Statistics Theory (math.ST); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
- [10] arXiv:2212.10717 [pdf, other]
-
Title: Hidden Poison: Machine Unlearning Enables Camouflaged Poisoning AttacksSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computers and Society (cs.CY)
- [11] arXiv:2206.02617 [pdf, other]
-
Title: Individual Privacy Accounting for Differentially Private Stochastic Gradient DescentComments: Add clarification about the applicability of Definition 4Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
- [12] arXiv:2405.20405 [pdf, other]
-
Title: Private Mean Estimation with Person-Level Differential PrivacyAuthors: Sushant Agarwal, Gautam Kamath, Mahbod Majid, Argyris Mouzakis, Rose Silver, Jonathan UllmanComments: 72 pages, 3 figuresSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [13] arXiv:2212.06470 [pdf, ps, other]
-
Title: Position: Considerations for Differentially Private Learning with Large-Scale Public PretrainingComments: Full and unabridged version of paper ICML 2024Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [14] arXiv:2406.17814 [pdf, ps, other]
-
Title: Distribution Learnability and RobustnessComments: In NeurIPS 2023Subjects: Machine Learning (stat.ML); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [15] arXiv:2212.05015 [pdf, ps, other]
-
Title: Robustness Implies Privacy in Statistical EstimationComments: 90 pages, 2 tables. Appeared in STOC, 2023Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (stat.ML)
- [16] arXiv:2404.06737 [pdf, other]
-
Title: Disguised Copyright Infringement of Latent Diffusion ModelsComments: Accepted to ICML 2024Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
- [17] arXiv:2405.04034 [pdf, other]
-
Title: Differentially Private Post-Processing for Fair RegressionComments: ICML 2024. Code is at this https URLSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Computers and Society (cs.CY)
- [18] arXiv:2304.06929 [pdf, ps, other]
-
Title: Advancing Differential Privacy: Where We Are Now and Future Directions for Real-World DeploymentAuthors: Rachel Cummings, Damien Desfontaines, David Evans, Roxana Geambasu, Yangsibo Huang, Matthew Jagielski, Peter Kairouz, Gautam Kamath, Sewoong Oh, Olga Ohrimenko, Nicolas Papernot, Ryan Rogers, Milan Shen, Shuang Song, Weijie Su, Andreas Terzis, Abhradeep Thakurta, Sergei Vassilvitskii, Yu-Xiang Wang, Li Xiong, Sergey Yekhanin, Da Yu, Huanyu Zhang, Wanrong ZhangSubjects: Cryptography and Security (cs.CR)
- [19] arXiv:2402.12626 [pdf, other]
-
Title: Indiscriminate Data Poisoning Attacks on Pre-trained Feature ExtractorsComments: Accepted to SaTML 2024Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
- [20] arXiv:2204.09092 [pdf, other]
-
Title: Indiscriminate Data Poisoning Attacks on Neural NetworksComments: Accepted to TMLR in 2022Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
- [21] arXiv:2110.14465 [pdf, other]
-
Title: Unbiased Statistical Estimation and Valid Confidence Intervals Under Differential PrivacySubjects: Methodology (stat.ME); Cryptography and Security (cs.CR); Statistics Theory (math.ST)
- [22] arXiv:2311.06477 [pdf, other]
-
Title: Report of the 1st Workshop on Generative AI and LawAuthors: A. Feder Cooper, Katherine Lee, James Grimmelmann, Daphne Ippolito, Christopher Callison-Burch, Christopher A. Choquette-Choo, Niloofar Mireshghallah, Miles Brundage, David Mimno, Madiha Zahrah Choksi, Jack M. Balkin, Nicholas Carlini, Christopher De Sa, Jonathan Frankle, Deep Ganguli, Bryant Gipson, Andres Guadamuz, Swee Leng Harris, Abigail Z. Jacobs, Elizabeth Joh, Gautam Kamath, Mark Lemley, Cass Matthews, Christine McLeavey, Corynne McSherry, Milad Nasr, Paul Ohm, Adam Roberts, Tom Rubin, Pamela Samuelson, Ludwig Schubert, Kristen Vaccaro, Luis Villa, Felix Wu, Elana ZeideSubjects: Computers and Society (cs.CY)
- [23] arXiv:2302.02936 [pdf, other]
-
Title: Private GANs, RevisitedComments: 28 pages; revisions and new experiments from TMLR camera-ready + code release at this https URLSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
- [24] arXiv:2308.06239 [pdf, ps, other]
-
Title: Private Distribution Learning with Public Data: The View from Sample CompressionComments: 31 pagesSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [25] arXiv:2303.03592 [pdf, other]
-
Title: Exploring the Limits of Model-Targeted Indiscriminate Data Poisoning AttacksComments: Accepted to ICML 2023Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
- [26] arXiv:2208.07984 [pdf, other]
-
Title: Private Estimation with Public DataComments: 55 pages; updated funding acknowledgement + simulation results from NeurIPS 2022 camera-readySubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [27] arXiv:2205.08532 [pdf, ps, other]
-
Title: New Lower Bounds for Private Estimation and a Generalized Fingerprinting LemmaComments: NeurIPS 2022. Minor correction to the discussion of independent workSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [28] arXiv:2303.01256 [pdf, other]
-
Title: Choosing Public Datasets for Private Machine Learning via Gradient Subspace DistanceSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
- [29] arXiv:2106.01336 [pdf, ps, other]
-
Title: Improved Rates for Differentially Private Stochastic Convex Optimization with Heavy-Tailed DataSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Optimization and Control (math.OC); Machine Learning (stat.ML)
- [30] arXiv:2006.06618 [pdf, other]
-
Title: CoinPress: Practical Private Mean and Covariance EstimationComments: Code is available at this https URLSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [31] arXiv:2110.06500 [pdf, other]
-
Title: Differentially Private Fine-tuning of Language ModelsAuthors: Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi, Huseyin A. Inan, Gautam Kamath, Janardhan Kulkarni, Yin Tat Lee, Andre Manoel, Lukas Wutschitz, Sergey Yekhanin, Huishuai ZhangComments: ICLR 2022. Code available at this https URLSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
- [32] arXiv:2111.12981 [pdf, ps, other]
-
Title: Efficient Mean Estimation with Pure Differential Privacy via a Sum-of-Squares Exponential MechanismComments: 66 pages, STOC 2022Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (stat.ML)
- [33] arXiv:1905.11947 [pdf, ps, other]
-
Title: Private Identity Testing for High-Dimensional DistributionsComments: Discussing a mistake in the proof of one of the algorithms (Theorem 1.2, computationally inefficient tester), and pointing to follow-up work by Narayanan (2022) who improves upon our results and fixes this mistakeSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [34] arXiv:2111.05320 [pdf, ps, other]
-
Title: Robust Estimation for Random GraphsSubjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Statistics Theory (math.ST); Machine Learning (stat.ML)
- [35] arXiv:2111.04609 [pdf, ps, other]
-
Title: A Private and Computationally-Efficient Estimator for Unbounded GaussiansSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG)
- [36] arXiv:2201.11308 [pdf, other]
-
Title: Calibration with Privacy in Peer ReviewComments: 31 pages, 6 figuresSubjects: Cryptography and Security (cs.CR); Information Theory (cs.IT)
- [37] arXiv:2111.04906 [pdf, other]
-
Title: The Role of Adaptive Optimizers for Honest Private Hyperparameter SelectionSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
- [38] arXiv:2106.13414 [pdf, other]
-
Title: The Price of Tolerance in Distribution TestingComments: Added a result on instance-optimal testing, and further discussion in the introductionSubjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Probability (math.PR); Statistics Theory (math.ST); Machine Learning (stat.ML)
- [39] arXiv:2010.09063 [pdf, other]
-
Title: Enabling Fast Differentially Private SGD via Just-in-Time Compilation and VectorizationComments: To appear in NeurIPS 2021Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Performance (cs.PF)
- [40] arXiv:2103.03279 [pdf, ps, other]
-
Title: Remember What You Want to Forget: Algorithms for Machine UnlearningSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
- [41] arXiv:2002.09464 [pdf, other]
-
Title: Private Mean Estimation of Heavy-Tailed DistributionsComments: Appeared in COLT 2020Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [42] arXiv:1911.07357 [pdf, ps, other]
-
Title: Random Restrictions of High-Dimensional Distributions and Uniformity Testing with Subcube ConditioningComments: Added Remark 4.4, which discusses the time complexity (the algorithms are polynomial-time, based on an observation from [CJLW20]); removing log log log n factor for the Gaussian testing algorithm. These changes reflect those included in the conference version (SODA'21)Subjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [43] arXiv:1905.13229 [pdf, ps, other]
-
Title: Private Hypothesis SelectionComments: Appeared in NeurIPS 2019. Final version to appear in IEEE Transactions on Information TheorySubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [44] arXiv:2002.12321 [pdf, other]
-
Title: PAPRIKA: Private Online False Discovery Rate ControlSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Statistics Theory (math.ST); Methodology (stat.ME)
- [45] arXiv:2010.09929 [pdf, ps, other]
-
Title: On the Sample Complexity of Privately Learning Unbounded High-Dimensional GaussiansSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG)
- [46] arXiv:2002.09463 [pdf, ps, other]
-
Title: Privately Learning Markov Random FieldsSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [47] arXiv:2002.09465 [pdf, other]
-
Title: Locally Private Hypothesis SelectionAuthors: Sivakanth Gopi, Gautam Kamath, Janardhan Kulkarni, Aleksandar Nikolov, Zhiwei Steven Wu, Huanyu ZhangComments: To appear in COLT 2020Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [48] arXiv:2005.00010 [pdf, other]
-
Title: A Primer on Private StatisticsComments: 20 pages. Comments welcomeSubjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG)
- [49] arXiv:1909.03951 [pdf, other]
-
Title: Differentially Private Algorithms for Learning Mixtures of Separated GaussiansComments: To appear in NeurIPS 2019Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [50] arXiv:1612.03147 [pdf, ps, other]
-
Title: Testing Ising ModelsComments: Appeared SODA 2018. Final version to appear in IEEE Transactions on Information TheorySubjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [51] arXiv:1805.00216 [pdf, other]
-
Title: Privately Learning High-Dimensional DistributionsComments: To appear in COLT 2019Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [52] arXiv:1803.02815 [pdf, other]
-
Title: Sever: A Robust Meta-Algorithm for Stochastic OptimizationAuthors: Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Jacob Steinhardt, Alistair StewartComments: To appear in ICML 2019Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
- [53] arXiv:1811.11148 [pdf, ps, other]
-
Title: The Structure of Optimal Private Tests for Simple HypothesesComments: To appear in STOC 2019Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
- [54] arXiv:1604.06443 [pdf, ps, other]
-
Title: Robust Estimators in High Dimensions without the Computational IntractabilitySubjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
- [55] arXiv:1411.7346 [pdf, ps, other]
-
Title: A Chasm Between Identity and Equivalence Testing with Conditional QueriesComments: 39 pages. To appear in Theory of Computing. Preliminary version appeared in RANDOM 2015Subjects: Data Structures and Algorithms (cs.DS); Computational Complexity (cs.CC); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [56] arXiv:1807.06168 [pdf, ps, other]
-
Title: Anaconda: A Non-Adaptive Conditional Sampling Algorithm for Distribution TestingComments: SODA 2019Subjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [57] arXiv:1703.00893 [pdf, other]
-
Title: Being Robust (in High Dimensions) Can Be PracticalAuthors: Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Ankur Moitra, Alistair StewartComments: Appeared in ICML 2017Subjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (stat.ML)
- [58] arXiv:1803.00008 [pdf, other]
-
Title: INSPECTRE: Privately Estimating the UnseenSubjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [59] arXiv:1802.07229 [pdf, other]
-
Title: Actively Avoiding Nonsense in Generative ModelsSubjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
- [60] arXiv:1704.03866 [pdf, ps, other]
-
Title: Robustly Learning a Gaussian: Getting Optimal Error, EfficientlyAuthors: Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Ankur Moitra, Alistair StewartComments: To appear in SODA 2018Subjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
- [61] arXiv:1708.00002 [pdf, ps, other]
-
Title: Which Distribution Distances are Sublinearly Testable?Comments: To appear in SODA 2018Subjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [62] arXiv:1710.04170 [pdf, ps, other]
-
Title: Concentration of Multilinear Functions of the Ising Model with Applications to Network DataComments: To appear in NIPS 2017Subjects: Probability (math.PR); Machine Learning (cs.LG); Mathematical Physics (math-ph); Statistics Theory (math.ST); Machine Learning (stat.ML)
- [63] arXiv:1703.10127 [pdf, other]
-
Title: Priv'IT: Private and Sample Efficient Identity TestingComments: To appear in ICML 2017Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [64] arXiv:1511.03641 [pdf, ps, other]
-
Title: A Size-Free CLT for Poisson Multinomials and its ApplicationsComments: To appear in STOC 2016Subjects: Data Structures and Algorithms (cs.DS); Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [65] arXiv:1507.05952 [pdf, ps, other]
-
Title: Optimal Testing for Properties of DistributionsComments: 31 pages, extended abstract appeared as a spotlight in NIPS 2015Subjects: Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Machine Learning (cs.LG); Statistics Theory (math.ST)
- [66] arXiv:1504.08363 [pdf, ps, other]
-
Title: On the Structure, Covering, and Learning of Poisson Multinomial DistributionsComments: 49 pages, extended abstract appeared in FOCS 2015Subjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [67] arXiv:1312.1054 [pdf, ps, other]
-
Title: Faster and Sample Near-Optimal Algorithms for Proper Learning Mixtures of GaussiansComments: 31 pages, to appear in COLT 2014Subjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
- [68] arXiv:1203.6346 [pdf, other]
-
Title: An Analysis of One-Dimensional Schelling SegregationComments: 24 pages, to appear in STOC 2012Subjects: Computer Science and Game Theory (cs.GT)
The web address for this page and the arXiv author id for Gautam Kamath is http://arxiv.org/a/kamath_g_1. There is also an Atom feed available from http://arxiv.org/a/kamath_g_1.atom2 (authors combined, best for most current feed readers), and http://arxiv.org/a/kamath_g_1.atom (authors in separate atom:author elements).
See author identifier help for more information about arXiv author identifiers, please report any problems.