Gautam Kamath

Office: DC 1311

Cell Phone: 657-c0mpsci (657-206-7724)

Office Phone: 519-888-4567 ext. 31245

Email: g@csail.mit.edu

Links to: CV (as of July 2024); Google Scholar; DBLP; arXiv; GitHub; Twitter; Youtube; Blog.

I host office hours as part of the Learning Theory Alliance. Book a slot here.

Feel free to send me comments anonymously here. Note that, by design, I won't be able to respond -- if you need me to, it's better handled via email, or leaving some way for me to reply.

I am unable to respond to requests for internships or graduate admission applications from outside the university. If you do feel compelled to contact me in this way, I am far more likely to remember you if you demonstrate that you have genuinely engaged with my work. Instead, I recommend that you apply for graduate studies through the standard channels.

About Me

I am an Assistant Professor at the University of Waterloo's Cheriton School of Computer Science, a Faculty Member at the Vector Institute, and a Canada CIFAR AI Chair. I run The Salon. I'm interested in reliable and trustworthy statistics and machine learning, including considerations such as data privacy and robustness.

I was a Microsoft Research Fellow at the Simons Institute for the Theory of Computing for the Fall 2018 semester program on Foundations of Data Science and the Spring 2019 semester program on Data Privacy: Foundations and Applications. Before that, I completed my Ph.D. at MIT, affiliated with the Theory of Computing group in CSAIL. I was very fortunate to be advised by Costis Daskalakis.

Before MIT, I spent four wonderful years at Cornell University, graduating in May 2012 with a degree in Computer Science and Electrical and Computer Engineering. At Cornell, I was incredibly lucky to have the opportunity to work with Bobby Kleinberg.

News

(5/14/25) Mentee Valentio Iverson wins the Germain-Erdős Undergraduate Award in Mathematical Research.
(5/3/25) Optimal Differentially Private Sampling of Unbounded Gaussians accepted to COLT 2025.
(5/1/25) On the Learnability of Distribution Classes with Adaptive Adversaries accepted to ICML 2025.
(4/29/25) I am a recipient of the Ontario Early Researcher Award.
(4/9/25) I am joining the editorial board of the Journal of Privacy and Confidentiality.
(3/3/25) New paper on arXiv: Optimal Differentially Private Sampling of Unbounded Gaussians.
(2/28/25) I will be on the program committee for SODA 2026.
(2/5/25) Quoted in a Toronto Star article about data privacy.
(1/22/25) One paper accepted to ICLR 2025.
(1/21/25) I will be speaking at Symposium on Mathematical Foundations of Trustworthy Machine Learning in Ascona, Switzerland in October 2025.
(1/21/25) I will be participating in an ACL Mentorship Panel.
(1/14/25) I was quoted in a news article about NeurIPS high school projects.
(1/2/25) Undergraduate advisee Chris Trevisan wins the 2025 CRA Outstanding Undergraduate Researcher Award.
(1/1/25) Unbiased Statistical Estimation and Valid Confidence Intervals Under Differential Privacy published in Statistica Sinica.
(12/30/24) New paper on arXiv: BridgePure: Revealing the Fragility of Black-box Data Protection.
(12/16/24) I will be an AC for the ICML 2025 Position Paper track.
(12/13/24) Three papers (1, 2, 3) accepted to SaTML 2025.
(12/11/24) A few trips to Italy planned for early 2025: ALT 2025 in Milan, and an ELSA workshop on privacy in Bertinoro.
(12/10/24) A Bias-Accuracy-Privacy Trilemma for Statistical Estimation accepted to the Journal of the American Statistical Association.
(12/3/24) I wrote a survey on connections between various types of robustness, including contamination, heavy-tailed data, and privacy.
(11/15/24) I will be giving talks at two Vector Institute events this month: ICML and ICLR 2024 Conference Highlights, and ML Theory Day.
(11/8/24) I will be a Senior PC member for COLT 2025.
(11/7/24) I will be giving a talk at the 2024 Charles River Privacy Day.
(10/04/24) Private Mean Estimation with Person-Level Differential Privacy accepted to SODA 2025.
(9/30/24) New paper on arXiv: Membership Inference Attacks Cannot Prove that a Model Was Trained On Your Data.
(9/16/24) I will be attending the Encore DP workshop in January 2025.
(9/9/24) I will be giving talks at Google's Core Labs seminar on September 25 and NUSAiL AI Horizons seminar on October 3.
(9/2/24) I will be on Ruicheng Xian's PhD committee.
(8/24/24) I will be on the PC of STOC 2025.
(8/20/24) I will be doing an online AMA at the IEEE Computer Society Santa Clara Valley Chapter for IEEE Day.
(7/29/24) I will be a reader for Linfeng Ye's Master's thesis.
(7/25/24) I was quoted in an MIT Technology Review article on copyright traps.
(7/23/24) "Position: Considerations for Differentially Private Learning with Large-Scale Public Pretraining" recognized as a Best Paper at ICML 2024.
(7/18/24) "The Discrete Gaussian for Differential Privacy" recognized with the 2024 Caspar Bowden Award for Outstanding Research in Privacy Enhancing Technologies.
(7/1/24) I am a director of the Association for Computational Learning.
(7/1/24) I'm giving the opening keynote at the Toronto Machine Learning Summit.
(6/24/24) New paper on arXiv.
(6/2/24) Two new papers (1, 2) on arXiv.
(5/30/24) I will be program committee co-chair for ALT 2025.
(5/28/24) I gave a talk at the 2024 Canadian Computing Olympiad.
(5/13/24) I am co-organizing a robustness workshop at TTIC from June 12 to 14.
(5/3/24) I will be the social media chair for SaTML 2025.
(5/1/24) Three papers (1, 2, 3) accepted to ICML 2024.
(4/12/24) I will be on the PC of TPDP 2024.
(4/11/24) New paper on arXiv: Disguised Copyright Infringement of Latent Diffusion Models.
(4/2/24) I will be on the PhD thesis committee of Aseem Baranwal.
(4/1/24) I gave a (virtual) guest lecture at Rex Ying's Trustworthy Deep Learning course at Yale.
(3/28/24) I will be an Area Chair for NeurIPS 2024.
(3/7/24) I was quoted in a couple of articles (1, 2) on protections for artists against ML model training.

Older News

Research

Most authorships are in alphabetical order, as is common in areas of computer science. Papers with contribution-order authorship are indicated, and equal contributions are marked with * or ^. Generally, these will put the students as first-author, with equal contribution amongst the senior authors. ⓡ is used for randomized author order.

Selected Papers (Show all)

BridgePure: Revealing the Fragility of Black-box Data Protection.
Yihan Wang*, Yiwei Lu*, Xiao-Shan Gao^, Gautam Kamath^, Yaoliang Yu^. (Contribution order)
Manuscript.

The Broader Landscape of Robustness in Algorithmic Statistics.
Gautam Kamath.
Manuscript.

Optimal Differentially Private Sampling of Unbounded Gaussians.
Valentio Iverson, Gautam Kamath, Argyris Mouzakis.
Proceedings of the 38th Annual Conference on Learning Theory (COLT 2025).

On the Learnability of Distribution Classes with Adaptive Adversaries.
Tosca Lechner, Alex Bie*, Gautam Kamath*. (Contribution order)
Proceedings of the 42nd International Conference on Machine Learning (ICML 2025).

Machine Unlearning Fails to Remove Data Poisoning Attacks.
Martin Pawelczyk*, Jimmy Z. Di*, Yiwei Lu, Gautam Kamath^, Ayush Sekhari^, Seth Neel^. (Contribution order)
Proceedings of the 13th International Conference on Learning Representations (ICLR 2025).

Presented at the ICML 2024 2nd Workshop on Generative AI and Law. Oral Presentation.

A Bias-Accuracy-Privacy Trilemma for Statistical Estimation.
Gautam Kamath, Argyris Mouzakis, Matthew Regehr, Vikrant Singhal, Thomas Steinke, Jonathan Ullman.
Journal of the American Statistical Association, to appear, 2025+.

Presented at TPDP 2023.

Membership Inference Attacks Cannot Prove that a Model Was Trained On Your Data.
Jie Zhang, Debeshee Das, Gautam Kamath, Florian Tramèr. (Contribution order)
Proceedings of the 2025 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML 2025).

Avoiding Pitfalls for Privacy Accounting of Subsampled Mechanisms under Composition.
Christian Janos Lebeda*, Matthew Regehr*, Gautam Kamath^, Thomas Steinke^. (Contribution order)
Proceedings of the 2025 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML 2025).

Choosing Public Datasets for Private Machine Learning via Gradient Subspace Distance.
Xin Gu, Gautam Kamath*, Zhiwei Steven Wu*. (Contribution order)
Proceedings of the 2025 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML 2025).

Presented at the NeurIPS 2022 Workshop on Distribution Shifts.
Presented at TPDP 2023.

Private Mean Estimation with Person-Level Differential Privacy.
Sushant Agarwal, Gautam Kamath, Mahbod Majid, Argyris Mouzakis, Rose Silver, Jonathan Ullman.
Proceedings of the 36th Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2025).

Unbiased Statistical Estimation and Valid Confidence Intervals Under Differential Privacy.
Christian Covington, Xi He*, James Honaker*, Gautam Kamath*. (Contribution order)
Statistica Sinica, 35(Online Special Issue I), 2025.

Presented at the ICML 2021 Workshop on Theory and Practice of Differential Privacy (TPDP 2021).

Differentially Private Post-Processing for Fair Regression.
Ruicheng Xian, Qiaobo Li, Gautam Kamath, Han Zhao (Contribution order)
Proceedings of the 41st International Conference on Machine Learning (ICML 2024).

Disguised Copyright Infringement of Latent Diffusion Models.
Yiwei Lu*, Matthew Y.R. Yang*, Zuoqiu Liu*, Gautam Kamath^, Yaoliang Yu^. (Contribution order)
Proceedings of the 41st International Conference on Machine Learning (ICML 2024).

Presented at the ICML 2024 2nd Workshop on Generative AI and Law.

Position: Considerations for Differentially Private Learning with Large-Scale Public Pretraining.
Florian Tramèr*, Gautam Kamath*, Nicholas Carlini*. (Reverse alphabetical order)
Proceedings of the 41st International Conference on Machine Learning (ICML 2024). Oral Presentation

ICML 2024 Best Paper Award.
Presented at TPDP 2023.

Indiscriminate Data Poisoning Attacks on Pre-trained Feature Extractors.
Yiwei Lu, Matthew Y.R. Yang, Gautam Kamath*, Yaoliang Yu*. (Contribution order)
Proceedings of the 2024 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML 2024).

Not All Learnable Distribution Classes are Privately Learnable.
Mark Bun, Gautam Kamath, Argyris Mouzakis, Vikrant Singhal.
Proceedings of the 35th International Conference on Algorithmic Learning Theory (ALT 2024).

Distribution Learnability and Robustness.
Shai Ben-David, Alex Bie, Gautam Kamath, Tosca Lechner.
Advances in Neural Information Processing Systems 36 (NeurIPS 2023).

Private Distribution Learning with Public Data: The View from Sample Compression.
Shai Ben-David, Alex Bie, Clément L. Canonne, Gautam Kamath, Vikrant Singhal.
Advances in Neural Information Processing Systems 36 (NeurIPS 2023).

Selected for a Spotlight Presentation.
Presented at TPDP 2023.

Hidden Poison: Machine Unlearning Enables Camouflaged Poisoning Attacks.
Jimmy Z. Di, Jack Douglas, Jayadev Acharya*, Gautam Kamath*, Ayush Sekhari*. (Contribution order)
Advances in Neural Information Processing Systems 36 (NeurIPS 2023).

Presented at the NeurIPS 2022 ML Safety Workshop.
Presented at the NeurIPS 2022 workshop on Trustworthy and Socially Responsible Machine Learning.

Private GANs, Revisited.
Alex Bie, Gautam Kamath*, Guojun Zhang*. (Contribution order)
Transactions on Machine Learning Research (TMLR), 2023. Survey Certification.

Presented at the NeurIPS 2022 SyntheticData4ML Workshop.
Presented at TPDP 2023.

Individual Privacy Accounting for Differentially Private Stochastic Gradient Descent.
Da Yu, Gautam Kamath*, Janardhan Kulkarni*, Tie-Yan Liu*, Jian Yin*, Huishuai Zhang*. (Contribution order)
Transactions on Machine Learning Research (TMLR), 2023.

Exploring the Limits of Model-Targeted Indiscriminate Data Poisoning Attacks.
Yiwei Lu, Gautam Kamath*, Yaoliang Yu*. (Contribution order)
Proceedings of the 40th International Conference on Machine Learning (ICML 2023).

Robustness Implies Privacy in Statistical Estimation.
Samuel B. Hopkins, Gautam Kamath, Mahbod Majid, Shyam Narayanan.
Proceedings of the 55th ACM Symposium on Theory of Computing (STOC 2023).

Presented at TPDP 2023. Oral Presentation.
Sam gave an invited survey on this line of work at HALG 2024.

Indiscriminate Data Poisoning Attacks on Neural Networks.
Yiwei Lu, Gautam Kamath*, Yaoliang Yu*. (Contribution order)
Transactions on Machine Learning Research (TMLR), 2022.

Presented at the NeurIPS 2022 ML Safety Workshop.
Presented at the NeurIPS 2022 workshop on Trustworthy and Socially Responsible Machine Learning.

New Lower Bounds for Private Estimation and a Generalized Fingerprinting Lemma.
Gautam Kamath, Argyris Mouzakis, Vikrant Singhal.
Advances in Neural Information Processing Systems 35 (NeurIPS 2022).

Presented at the ICML 2022 Workshop on Theory and Practice of Differential Privacy (TPDP 2022).

Private Estimation with Public Data.
Alex Bie, Gautam Kamath, Vikrant Singhal.
Advances in Neural Information Processing Systems 35 (NeurIPS 2022).

Presented at the ICML 2022 Workshop on Theory and Practice of Differential Privacy (TPDP 2022).

Improved Rates for Differentially Private Stochastic Convex Optimization with Heavy-Tailed Data.
Gautam Kamath, Xingtu Liu, Huanyu Zhang.
Proceedings of the 39th International Conference on Machine Learning (ICML 2022).

Selected for a long presentation. (2.1% acceptance rate).
Presented at the ICML 2021 Workshop on Theory and Practice of Differential Privacy (TPDP 2021).

Robust Estimation for Random Graphs.
Jayadev Acharya, Ayush Jain, Gautam Kamath, Ananda Theertha Suresh, Huanyu Zhang.
Proceedings of the 35th Annual Conference on Learning Theory (COLT 2022).

A Private and Computationally-Efficient Estimator for Unbounded Gaussians.
Gautam Kamath, Argyris Mouzakis, Vikrant Singhal, Thomas Steinke, Jonathan Ullman.
Proceedings of the 35th Annual Conference on Learning Theory (COLT 2022).

Presented at the ICML 2022 Workshop on Theory and Practice of Differential Privacy (TPDP 2022).

The Price of Tolerance in Distribution Testing.
Clément L. Canonne, Ayush Jain, Gautam Kamath, Jerry Li.
Proceedings of the 35th Annual Conference on Learning Theory (COLT 2022).

Calibration with Privacy in Peer Review.
Wenxin Ding, Gautam Kamath ⓡ Weina Wang ⓡ Nihar B. Shah.
Proceedings of the 2022 IEEE International Symposium on Information Theory (ISIT 2022).

Presented at the AAAI 2022 Workshop on Privacy-Preserving Artificial Intelligence (PPAI 2022). Oral Presentation.

Efficient Mean Estimation with Pure Differential Privacy via a Sum-of-Squares Exponential Mechanism.
Samuel B. Hopkins, Gautam Kamath, Mahbod Majid.
Proceedings of the 54th ACM Symposium on Theory of Computing (STOC 2022).

University of Waterloo Faculty of Mathematics Graduate Research Excellence Award.
Presented at FORC 2022, non-archival track.
Presented at the ICML 2022 Workshop on Theory and Practice of Differential Privacy (TPDP 2022).

Differentially Private Fine-tuning of Language Models.
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 Zhang*. (Contribution order)
Journal of Privacy and Confidentiality, 14(2), 2024.
Proceedings of the 10th International Conference on Learning Representations (ICLR 2022).

Featured in: Microsoft Research Blog.
Presented at the ICML 2022 Workshop on Theory and Practice of Differential Privacy (TPDP 2022).

The Role of Adaptive Optimizers for Honest Private Hyperparameter Selection.
Shubhankar Mohapatra*, Sajin Sasy*, Xi He^, Gautam Kamath^, Om Thakkar^. (Contribution order)
Proceedings of the Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI 2022).

Selected for an Oral Presentation (4.3% acceptance rate).
Presented at the ICML 2021 Workshop on Theory and Practice of Differential Privacy (TPDP 2021).

Remember What You Want to Forget: Algorithms for Machine Unlearning.
Ayush Sekhari, Jayadev Acharya*, Gautam Kamath*, Ananda Theertha Suresh*. (Contribution order)
Advances in Neural Information Processing Systems 34 (NeurIPS 2021).

Presented at the ICML 2021 Workshop on Theory and Practice of Differential Privacy (TPDP 2021).
Featured in: WIRED Magazine.

Enabling Fast Differentially Private SGD via Just-in-Time Compilation and Vectorization.
Pranav Subramani*, Nicholas Vadivelu*, Gautam Kamath. (Contribution order)
Advances in Neural Information Processing Systems 34 (NeurIPS 2021).

Presented at the NeurIPS 2020 Workshop on Privacy Preserving Machine Learning (PPML 2020).
Presented at the ICML 2021 Workshop on Theory and Practice of Differential Privacy (TPDP 2021).
Code is available here.
Inspired changes in TensorFlow Privacy.
Featured in: DeepMind Blog (Using JAX to accelerate our research), Google AI Blog (Google Research: Looking Back at 2020, and Forward to 2021).

PAPRIKA: Private Online False Discovery Rate Control.
Wanrong Zhang, Gautam Kamath*, Rachel Cummings*. (Contribution order)
Proceedings of the 38th International Conference on Machine Learning (ICML 2021).

Presented at FORC 2021, non-archival track.
Presented at the CCS 2020 Workshop on Theory and Practice of Differential Privacy (TPDP 2020).
Code is available here.

On the Sample Complexity of Privately Learning Unbounded High-Dimensional Gaussians.
Ishaq Aden-Ali, Hassan Ashtiani, Gautam Kamath.
Proceedings of the 32nd International Conference on Algorithmic Learning Theory (ALT 2021).

Presented at the NeurIPS 2020 Workshop on Privacy Preserving Machine Learning (PPML 2020).

Random Restrictions of High-Dimensional Distributions and Uniformity Testing with Subcube Conditioning.
Clément L. Canonne, Xi Chen, Gautam Kamath, Amit Levi, Erik Waingarten.
Proceedings of the 32nd Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2021).

ECCC link.
Featured in: Property Testing Review.
Talk by Erik here.

CoinPress: Practical Private Mean and Covariance Estimation.
Sourav Biswas, Yihe Dong, Gautam Kamath, Jonathan Ullman.
Advances in Neural Information Processing Systems 33 (NeurIPS 2020).

Presented at the CCS 2020 Workshop on Theory and Practice of Differential Privacy (TPDP 2020).
Video of a talk is available here.
Code is available here.

The Discrete Gaussian for Differential Privacy.
Clément L. Canonne, Gautam Kamath, Thomas Steinke.
Journal of Privacy and Confidentiality, 12(1), 2022.
Advances in Neural Information Processing Systems 33 (NeurIPS 2020).

Deployed in the 2020 US Census.
2024 Caspar Bowden Award for Outstanding Research in Privacy Enhancing Technologies.
Presented at the CCS 2020 Workshop on Theory and Practice of Differential Privacy (TPDP 2020).
Code is available here.

Private Identity Testing for High-Dimensional Distributions.
Clément L. Canonne, Gautam Kamath, Audra McMillan, Jonathan Ullman, Lydia Zakynthinou.
Advances in Neural Information Processing Systems 33 (NeurIPS 2020).

Selected for a Spotlight Presentation.
Presented at the CCS 2019 Workshop on Theory and Practice of Differential Privacy (TPDP 2019).
Presented at the NeurIPS 2019 Workshop on Privacy in Machine Learning (PriML 2019).
Featured in: Property Testing Review.

Privately Learning Markov Random Fields.
Huanyu Zhang, Gautam Kamath*, Janardhan Kulkarni*, Zhiwei Steven Wu*. (Contribution order)
Proceedings of the 37th International Conference on Machine Learning (ICML 2020).

Presented at the CCS 2019 Workshop on Theory and Practice of Differential Privacy (TPDP 2019).
Presented at the NeurIPS 2019 Workshop on Privacy in Machine Learning (PriML 2019).

Private Mean Estimation of Heavy-Tailed Distributions.
Gautam Kamath, Vikrant Singhal, Jonathan Ullman.
Proceedings of the 33rd Annual Conference on Learning Theory (COLT 2020).

Presented at the CCS 2020 Workshop on Theory and Practice of Differential Privacy (TPDP 2020).
COLT talk by Vikrant available here.

Locally Private Hypothesis Selection.
Sivakanth Gopi, Gautam Kamath, Janardhan Kulkarni, Aleksandar Nikolov, Zhiwei Steven Wu, Huanyu Zhang.
Proceedings of the 33rd Annual Conference on Learning Theory (COLT 2020).

Presented at the CCS 2020 Workshop on Theory and Practice of Differential Privacy (TPDP 2020).
Featured in: Property Testing Review.
I gave a talk more generally on hypothesis selection with privacy constraints: here.
COLT talk by Janardhan available here.

Differentially Private Algorithms for Learning Mixtures of Separated Gaussians.
Gautam Kamath, Or Sheffet, Vikrant Singhal, Jonathan Ullman.
Advances in Neural Information Processing Systems 32 (NeurIPS 2019).

Private Hypothesis Selection.
Mark Bun, Gautam Kamath, Thomas Steinke, Zhiwei Steven Wu.
IEEE Transactions on Information Theory, 67(3), 2021.
Advances in Neural Information Processing Systems 32 (NeurIPS 2019).

Presented at the CCS 2019 Workshop on Theory and Practice of Differential Privacy (TPDP 2019). Oral Presentation.
I gave a talk more generally on hypothesis selection with privacy constraints: here.
Video of a talk by Mark is available here.

Sever: A Robust Meta-Algorithm for Stochastic Optimization.
Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Jacob Steinhardt, Alistair Stewart.
Proceedings of the 36th International Conference on Machine Learning (ICML 2019).

Presented at the NeurIPS 2018 Workshop on Security in Machine Learning (SECML 2018). Oral Presentation.
Code is available here.
Video of a talk by Jerry is available here.

Privately Learning High-Dimensional Distributions.
Gautam Kamath, Jerry Li, Vikrant Singhal, Jonathan Ullman.
Proceedings of the 32nd Annual Conference on Learning Theory (COLT 2019).

Presented at the 2018 Workshop on Theory and Practice of Differential Privacy (TPDP 2018).
Video of a talk by Jon is available here.

The Structure of Optimal Private Tests for Simple Hypotheses.
Clément L. Canonne, Gautam Kamath, Audra McMillan, Adam Smith, Jonathan Ullman.
Proceedings of the 51st ACM Symposium on Theory of Computing (STOC 2019).

Video of a talk by Audra is available here.

Anaconda: A Non-Adaptive Conditional Sampling Algorithm for Distribution Testing.
Gautam Kamath, Christos Tzamos.
Proceedings of the 30th Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2019).

ECCC link.
Slides from my talk at SODA available here.
Featured in: Property Testing Review.

INSPECTRE: Privately Estimating the Unseen.
Jayadev Acharya, Gautam Kamath, Ziteng Sun, Huanyu Zhang.
Journal of Privacy and Confidentiality, 10(2), 2020. Special issue for TPDP 2018.
Proceedings of the 35th International Conference on Machine Learning (ICML 2018).

Presented at the CCS 2018 Workshop on Theory and Practice of Differential Privacy (TPDP 2018).
Code is available here.
Featured in: Property Testing Review
Video of a talk by Huanyu is available here.

Actively Avoiding Nonsense in Generative Models.
Steve Hanneke, Adam Kalai, Gautam Kamath, Christos Tzamos.
Proceedings of the 31st Annual Conference on Learning Theory (COLT 2018).

Video of Adam's talk at COLT can be viewed here.

Which Distribution Distances are Sublinearly Testable?
Constantinos Daskalakis, Gautam Kamath, John Wright.
Proceedings of the 29th Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2018).

ECCC link.
Featured in: Property Testing Review.

Testing Ising Models.
Constantinos Daskalakis, Nishanth Dikkala, Gautam Kamath.
IEEE Transactions on Information Theory, 65(11), 2019.
Proceedings of the 29th Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2018).

ECCC link.
Featured in: Property Testing Review.

Robustly Learning a Gaussian: Getting Optimal Error, Efficiently.
Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Ankur Moitra, Alistair Stewart.
Proceedings of the 29th Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2018).

Concentration of Multilinear Functions of the Ising Model with Applications to Network Data.
Constantinos Daskalakis, Nishanth Dikkala, Gautam Kamath.
Advances in Neural Information Processing Systems 30 (NIPS 2017).

Code is available here.
A three-minute summary video for NIPS is available here.
Our NIPS poster is viewable here.

Priv'IT: Private and Sample Efficient Identity Testing.
Bryan Cai, Constantinos Daskalakis, Gautam Kamath.
Proceedings of the 34th International Conference on Machine Learning (ICML 2017).

Presented at the ICML 2017 Workshop on Private and Secure Machine Learning 2017 (PSML 2017). Oral Presentation.
Video of my talk at ICML can be viewed here.
Code is available here.
Featured in: Property Testing Review.

Being Robust (in High Dimensions) Can Be Practical.
Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Ankur Moitra, Alistair Stewart.
Proceedings of the 34th International Conference on Machine Learning (ICML 2017).

Code is available here.
Video of Jerry's talk at ICML can be viewed here.

Robust Estimators in High Dimensions without the Computational Intractability.
Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Ankur Moitra, Alistair Stewart.
SIAM Journal on Computing, 48(2), 2019. Special Issue for FOCS 2016.
Proceedings of the 57th Annual IEEE Symposium on Foundations of Computer Science (FOCS 2016).

Invited to the SIAM Journal on Computing Special Issue for FOCS 2016.
Invited to Highlights of Algorithms 2017.
Invited to Communications of the ACM, Research Highlights.
Video of Jerry's talk at TCS+ can be viewed here. Video of Ankur's talk at the Simons Institute can be viewed here.
Featured in: MIT News, USC Viterbi News

A Size-Free CLT for Poisson Multinomials and its Applications.
Constantinos Daskalakis, Anindya De, Gautam Kamath, Christos Tzamos.
Proceedings of the 48th ACM Symposium on Theory of Computing (STOC 2016).

Optimal Testing for Properties of Distributions.
Jayadev Acharya, Constantinos Daskalakis, Gautam Kamath.
Advances in Neural Information Processing Systems 28 (NIPS 2015).

Selected for a Spotlight Presentation (4.5% acceptance rate).
Video of a talk by Costis at UT Austin can be viewed here. Video of the spotlight talk by Jayadev at NIPS can be viewed here (at 9:15).
Featured in: MIT Theory of Computation Student Blog, Property Testing Review.
An extremely condensed version of this paper appeared in Volume Four of Tiny Transactions on Computer Science (Tiny ToCS).

On the Structure, Covering, and Learning of Poisson Multinomial Distributions.
Constantinos Daskalakis, Gautam Kamath, Christos Tzamos.
Proceedings of the 56th Annual IEEE Symposium on Foundations of Computer Science (FOCS 2015).

A Chasm Between Identity and Equivalence Testing with Conditional Queries.
Jayadev Acharya, Clément L. Canonne, Gautam Kamath.
Theory of Computing, 14(19), 2018.
Proceedings of the 19th International Workshop on Randomization and Computation (RANDOM 2015).

ECCC link.
Slides from Clément's talk at RANDOM available here.
Featured in: Oded Goldreich's Choices, Property Testing Review, Sublinear.info.

Adaptive Estimation in Weighted Group Testing.
Jayadev Acharya, Clément L. Canonne, Gautam Kamath.
Proceedings of the 2015 IEEE International Symposium on Information Theory (ISIT 2015).

Faster and Sample Near-Optimal Algorithms for Proper Learning Mixtures of Gaussians.
Constantinos Daskalakis, Gautam Kamath.
Proceedings of the 27th Annual Conference on Learning Theory (COLT 2014).

Slides from the talk at COLT available here. Video of the talk can be viewed here.

An Analysis of One-Dimensional Schelling Segregation.
Christina Brandt, Nicole Immorlica, Gautam Kamath, Robert Kleinberg.
Proceedings of the 44th ACM Symposium on Theory of Computing (STOC 2012).

I was very lucky to be awarded the inaugural STOC Best Student Presentation Award by SIGACT for this talk.
Slides from the talk at STOC available here, adapted from slides from a talk by Nicole at MSR. Video of the talk can be viewed here.
Featured in: American Scientist.

Selected Workshop Papers (Show)

Machine Unlearning Fails to Remove Data Poisoning Attacks.
Martin Pawelczyk*, Jimmy Z. Di*, Yiwei Lu, Gautam Kamath^, Ayush Sekhari^, Seth Neel^. (Contribution order)
ICML 2024 2nd Workshop on Generative AI and Law. Oral Presentation.

Robustness Implies Privacy in Statistical Estimation.
Samuel B. Hopkins, Gautam Kamath, Mahbod Majid, Shyam Narayanan.
Theory and Practice of Differential Privacy 2023 (TPDP 2023). Oral Presentation.

Calibration with Privacy in Peer Review.
Wenxin Ding, Gautam Kamath ⓡ Weina Wang ⓡ Nihar B. Shah.
AAAI 2022 Workshop on Privacy-Preserving Artificial Intelligence (PPAI 2022). Oral Presentation.

Private Hypothesis Selection.
Mark Bun, Gautam Kamath, Thomas Steinke, Zhiwei Steven Wu.
CCS 2019 Workshop on Theory and Practice of Differential Privacy (TPDP 2019). Oral Presentation.

Sever: A Robust Meta-Algorithm for Stochastic Optimization.
Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Jacob Steinhardt, Alistair Stewart.
NeurIPS 2018 Workshop on Security in Machine Learning (SECML 2018). Oral Presentation.

Priv'IT: Private and Sample Efficient Identity Testing.
Bryan Cai, Constantinos Daskalakis, Gautam Kamath.
ICML 2017 Workshop on Private and Secure Machine Learning 2017 (PSML 2017). Oral Presentation.

Theses

Modern Challenges in Distribution Testing.
Gautam Kamath.
Ph.D. Thesis, 2018.
On Learning and Covering Structured Distributions.
Gautam Kamath.
M.Sc. Thesis, 2014.

Talk Videos

Here are some videos of talks I've given.

Hypothesis Selection with Privacy Constraints.
CoinPress: Practical Private Mean and Covariance Estimation.
Privately Learning High-Dimensional Distributions. Presented at Data Privacy: From Foundations to Applications.
Private Statistics: A TCS Perspective. Presented as part of the mini-course on private statistics at the Data Privacy Boot Camp.
Realizing Robustness. Presented at the Robust and High-Dimensional Statistics Workshop.
Modern Challenges in Distribution Testing. My PhD defense talk.
Differentially Private Hypothesis Testing and Property Estimation. Presented at Mathematical Foundations of Data Privacy.
Robustness in Unsupervised and Supervised Learning. Presented at Modern Challenges of Learning Theory.
Testing with Alternative Distances. Presented at Frontiers of Distribution Testing workshop at FOCS 2017.
Priv'IT: Private and Sample Efficient Identity Testing. Presented at ICML 2017.
Faster and Sample Near-Optimal Algorithms for Proper Learning Mixtures of Gaussians. Presented at COLT 2014.
An Analysis of One-Dimensional Schelling Segregation. Presented at STOC 2012.

Coauthors

My (150) co-authors include: Jayadev Acharya, Ishaq Aden-Ali, Sushant Agarwal, Hassan Ashtiani, Arturs Backurs, Jack M. Balkin, Shai Ben-David, Alex Bie, Sourav Biswas, Christina Brandt, Miles Brundage, Mark Bun, Bryan Cai, Christopher Callison-Burch, Clément L. Canonne, Nicholas Carlini, Rachel Cummings, Xi Chen, Madiha Zahrah Choksi, Christopher A. Choquette-Choo, A. Feder Cooper, Christian Covington, Debeshee Das, Constantinos Daskalakis, Anindya De, Christopher De Sa, Damien Desfontaines, Jimmy Z. Di, Ilias Diakonikolas, Nishanth Dikkala, Wenxin Ding, Yihe Dong, Jack Douglas, David Evans, Jonathan Frankle, Deep Ganguli, Xiao-Shan Gao, Roxana Geambasu, Bryant Gipson, Sivakanth Gopi, James Grimmelmann, Xin Gu, Andres Guadamuz, Steve Hanneke, Swee Leng Harris, Xi He, James Honaker, Samuel B. Hopkins, Yangsibo Huang, Nicole Immorlica, Huseyin A. Inan, Daphne Ippolito, Valentio Iverson, Abigail Z. Jacobs, Matthew Jagielski, Ayush Jain, Elizabeth Joh, Peter Kairouz, Adam Kalai, Daniel M. Kane, Robert Kleinberg, Janardhan Kulkarni, Christian Janos Lebeda, Tosca Lechner, Katherine Lee, Yin Tat Lee, Mark Lemley, Amit Levi, Jerry Li, Qiaobo Li, Tie-Yan Liu, Xingtu Liu, Zuoqiu Liu, Yiwei Lu, Mahbod Majid, Andre Manoel, Cass Matthews, Christine McLeavey, Corynne McSherry, David Mimno, Ankur Moitra, Shubhankar Mohapatra, Argyris Mouzakis, Audra McMillan, Niloofar Mireshghallah, Saurabh Naik, Shyam Narayanan, Milad Nasr, Seth Neel, Aleksandar Nikolov, Sewoong Oh, Paul Ohm, Olga Ohrimenko, Nicolas Papernot, Martin Pawelczyk, Matthew Regehr, Adam Roberts, Ryan Rogers, Tom Rubin, Pamela Samuelson, Sajin Sasy, Ludwig Schubert, Ayush Sekhari, Nihar B. Shah, Or Sheffet, Milan Shen, Rose Silver, Vikrant Singhal, Adam Smith, Shuang Song, Jacob Steinhardt, Thomas Steinke, Alistair Stewart, Weijie Su, Pranav Subramani, Ziteng Sun, Ananda Theertha Suresh, Andreas Terzis, Om Thakkar, Abhradeep Thakurta, Florian Tramèr, Christos Tzamos, Lukas Wutschitz, Jonathan Ullman, Kristen Vaccaro, Nicholas Vadivelu, Sergei Vassilvitskii, Luis Villa, Erik Waingarten, Yihan Wang, Weina Wang, Yu-Xiang Wang, John Wright, Felix Wu, Zhiwei Steven Wu, Ruicheng Xian, Li Xiong, Matthew Y.R. Yang, Sergey Yekhanin, Jian Yin, Da Yu, Yaoliang Yu, Lydia Zakynthinou, Elana Zeide, Guojun Zhang, Huanyu Zhang, Huishuai Zhang, Jie Zhang Wanrong Zhang, Han Zhao.

They originate from a number of countries, including Argentina, Australia, Austria, Brazil, Bulgaria, Canada, China, France, Germany, Greece, India, Indonesia, Iran, Israel, Latvia, Lebanon, New Zealand, Romania, Russia, South Korea, Switzerland, Turkey, United Kingdom, United States of America.

Other Writings

We wrote a "Behind the Screens" look at STOC 2020, a document which may be useful to other virtual conference organizers. With Konstantin Makarychev, Yury Makarychev, Madhur Tulsiani, and input from Julia Chuzhoy.
Jonathan Ullman and I wrote a brief survey of differentially private statistics: A Primer on Private Statistics.
For Advanced Algorithms (6.854), I wrote a survey of matroid secretary problem algorithms as a final project.
Some bounds on the expectation of the maximum of samples from a Gaussian. I thought this was a fairly natural problem, but when I wrote it, I was unable to find anyone else who had computed the explicit lower bound here.

Service

I am an Editor-in-Chief of TMLR.
I'm part of the Executive Committee of the Learning Theory Alliance.
I have been (or will be) program committee chair for the following conference: ALT 2025.
I have been (or will be) a general chair for the following conferences: STOC 2020.
I have been (or will be) program chair for the following workshops: TPDP 2021, TPDP 2022, UpML 2022.
I have been (or will be) on the core program committee or an area chair for the following conferences: SODA 2020, ICML 2020, ICALP 2020, Random 2020, ICLR 2021, FORC 2021, COLT 2021, CCS 2021, ESA 2021, NeurIPS 2021, SODA 2022, ICLR 2022, AAAI 2022, COLT 2022, NeurIPS 2022, SaTML 2023, FOCS 2023, ICLR 2023, ALT 2023, USENIX Security 2023, COLT 2023, FAccT 2023, ICML 2023, NeurIPS 2023, COLT 2024, NeurIPS 2024, STOC 2025, COLT 2025, ICML 2025 (Position Paper Track), SODA 2026.
I have been (or will be) on the program committee (i.e., a reviewer) of the following machine learning conferences: NIPS 2016, ICML 2018, NeurIPS 2018, AISTATS 2019, ICML 2019, NeurIPS 2019, AAAI 2020, AISTATS 2020, FAccT 2021, ALT 2021, ALT 2022, UAI 2022, CANAI 2024.
For completeness, conferences I have reviewed for include: AAAI, AISTATS, ALT, COLT, FAccT, FOCS, ICALP, ICML, ISAAC, ISIT, ITCS, NeurIPS, RANDOM, SODA, STACS, STOC.
I have (or will be) on the program committee of the following workshops: TPDP 2019, PriML 2019, TPDP 2020, PPML 2020, PriML 2021, ICBINB 2021, ICBINB 2022, ICBINB 2023, Regulatable ML 2023, PPAI 2024, TPDP 2024.
I mentor when I can. Some examples include at workshops for Women in Machine Learning, the Learning Theory Alliance, and junior-senior lunches at STOC/FOCS (which I sometimes organize). I also view my Twitter presence as a type of mentorship.
I am a maintainer of the CS Theory Blog Aggregator, along with Nima Anari and Arnab Bhattacharyya.
Clément Canonne and I organized a workshop called "A TCS Quiver" at FOCS 2019.
Clément Canonne and I organized a workshop on distribution testing at FOCS 2017.
Clément Canonne and I organized a workshop on orthogonal polynomials at FOCS 2016.
I was previously an editor for the MIT Theory of Computation Student Blog and Property Testing Review.
I'm one of the organizers of TCS+, an online seminar series in theoretical computer science, accessible to the widest possible audience, and ensuring a carbon-free dissemination of ideas across the globe.
I organized the second Sublinear Day, which was on April 10, 2015 at MIT.
I was the head organizer for the Second Annual Danny Lewin MIT Theory Student Retreat, which took place in October 2013. Aloni and Themis wrote a bit about this retreat here.
From Fall 2012 to Fall 2013, I was in charge of the Theory Group lunch, which was the current incarnation of Great Ideas in Theoretical Computer Science at CSAIL. The website for the current offering is here.

Other Undergrad Projects

As a final project for a microcontrollers class, my partner Dominick Grochowina and I made an automated vocalist for the game Rock Band. This was featured on the website Hack A Day. Videos of the bot in action can be seen here and here.
I worked on a project called Plannit, which won second place at Cornell Facebook Hackathon. Partners on this project were Jasdeep Hundal, Sohan Jain, and Harry Terkelsen.

Teaching

CS 480/680 - Introduction to Machine Learning: Winter 2021, Spring 2021, Spring 2022, Fall 2023
CS 860 - Algorithms for Private Data Analysis: Fall 2020, Fall 2022
CS 240 - Data Structures and Data Management: Spring 2020
CS 761 - Randomized Algorithms: Fall 2019

Advising

I'm very lucky to work with a number of great students -- check out their profiles on my group's People page.

Resources

Here is a collection of collections of talk videos.

TCS+: An series of online seminars in theoretical computer science.
Simons Institute Videos: Videos from workshops hosted at the Simons Institute for the Theory of Computing.
BIRS Videos: Videos from workshops hosted at the Banff International Research Station.
Institute for Advanced Studies Videos: Videos from the IAS. Note that many are related to other fields besides computer science.
Microsoft Research Talks: Talks at Microsoft Research, including a variety of topics beyond theory.
Shannon Channel: A series of online seminars in information theory.
Princeton TCS Videos: Videos from theory lunch and workshops within Princeton's theory group.
Videolectures.net: Lecture videos from a number of conferences and workshops, seems to be primarily focused on machine learning events.

Other

I used to go by the name "G", though I now prefer Gautam. Also, my name is not Guatam Kamath, though it is commonly misspelled as such.
My friends Aviv Adler, Greg Bodwin, and primarily Jennifer Tang, made a small puzzle hunt for my graduation. You can check out the puzzles here: 1, 2, 3, 4, 5, 6. Warning that they can be incredibly unfair if you are not me, but oh well. Try them out!