Volgen
Om Thakkar
Om Thakkar
Senior Research Scientist, Google
Geverifieerd e-mailadres voor google.com - Homepage
Titel
Geciteerd door
Geciteerd door
Jaar
Differentially private learning with adaptive clipping
G Andrew, O Thakkar, B McMahan, S Ramaswamy
Advances in Neural Information Processing Systems 34, 17455-17466, 2021
3012021
Towards Practical Differentially Private Convex Optimization
R Iyengar, JP Near, D Song, O Thakkar, A Thakurta, L Wang
IEEE Symposium on Security and Privacy 2019, 2019
1892019
Practical and private (deep) learning without sampling or shuffling
P Kairouz, B McMahan, S Song, O Thakkar, A Thakurta, Z Xu
International Conference on Machine Learning 2021, 5213-5225, 2021
1202021
Evading the curse of dimensionality in unconstrained private glms
S Song, T Steinke, O Thakkar, A Thakurta
International Conference on Artificial Intelligence and Statistics, 2638-2646, 2021
95*2021
Max-information, differential privacy, and post-selection hypothesis testing
R Rogers, A Roth, A Smith, O Thakkar
IEEE Symposium on Foundations of Computer Science 2016, 2016
892016
Model-Agnostic Private Learning
R Bassily, O Thakkar, A Thakurta
Neural Information Processing Systems 2018, 2018
84*2018
Privacy amplification via random check-ins
B Balle, P Kairouz, B McMahan, O Thakkar, A Guha Thakurta
Advances in Neural Information Processing Systems 33, 4623-4634, 2020
672020
Training production language models without memorizing user data
S Ramaswamy, O Thakkar, R Mathews, G Andrew, HB McMahan, ...
arXiv preprint arXiv:2009.10031, 2020
662020
Measuring forgetting of memorized training examples
M Jagielski, O Thakkar, F Tramer, D Ippolito, K Lee, N Carlini, E Wallace, ...
arXiv preprint arXiv:2207.00099, 2022
572022
Understanding unintended memorization in language models under federated learning
OD Thakkar, S Ramaswamy, R Mathews, F Beaufays
Proceedings of the Third Workshop on Privacy in Natural Language Processing …, 2021
52*2021
Public Data-Assisted Mirror Descent for Private Model Training
E Amid, A Ganesh, R Mathews, S Ramaswamy, S Song, T Steinke, ...
International Conference on Machine Learning 2022, 2021
462021
Differentially Private Matrix Completion, Revisited
P Jain, O Thakkar, A Thakurta
International Conference on Machine Learning 2018, 2018
422018
The Role of Adaptive Optimizers for Honest Private Hyperparameter Selection
S Mohapatra, S Sasy, X He, G Kamath, O Thakkar
36th AAAI Conference on Artificial Intelligence, 2021
312021
Revealing and protecting labels in distributed training
T Dang, O Thakkar, S Ramaswamy, R Mathews, P Chin, F Beaufays
Advances in Neural Information Processing Systems 34, 1727-1738, 2021
272021
Why is public pretraining necessary for private model training?
A Ganesh, M Haghifam, M Nasr, S Oh, T Steinke, O Thakkar, AG Thakurta, ...
International Conference on Machine Learning, 10611-10627, 2023
262023
Guaranteed validity for empirical approaches to adaptive data analysis
R Rogers, A Roth, A Smith, N Srebro, O Thakkar, B Woodworth
International Conference on Artificial Intelligence and Statistics, 2830-2840, 2020
102020
Detecting unintended memorization in language-model-fused ASR
WR Huang, S Chien, O Thakkar, R Mathews
arXiv preprint arXiv:2204.09606, 2022
92022
A method to reveal speaker identity in distributed asr training, and how to counter it
T Dang, O Thakkar, S Ramaswamy, R Mathews, P Chin, F Beaufays
ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and …, 2022
72022
Recycling scraps: Improving private learning by leveraging intermediate checkpoints
V Shejwalkar, A Ganesh, R Mathews, O Thakkar, A Thakurta
arXiv preprint arXiv:2210.01864, 2022
62022
Extracting targeted training data from ASR models, and how to mitigate it
E Amid, O Thakkar, A Narayanan, R Mathews, F Beaufays
arXiv preprint arXiv:2204.08345, 2022
42022
Het systeem kan de bewerking nu niet uitvoeren. Probeer het later opnieuw.
Artikelen 1–20