Differential privacy

Differentially private query release through adaptive projection

Sergul Aydore, William Brown, Michael Kearns, Krishnaram Kenthapadi, Luca Melis, Aaron Roth, Ankit Siva

ICML 2021, ICLR 2021 Workshop on Synthetic Data Generation

2021

We propose, implement, and evaluate a new algorithm for releasing answers to very large numbers of statistical queries like k-way marginals, subject to differential privacy. Our algorithm makes adaptive use of a continuous relaxation of the Projection Mechanism, which answers queries on the private dataset using simple perturbation, and then attempts to find the synthetic dataset that most closely matches

Security, privacy, and abuse prevention

Amazon Scholars Michael Kearns and Aaron Roth win 2021 PROSE Award

Staff writer

January 29, 2021

Authors’ book, The Ethical Algorithm, wins award in computer and information sciences category.

Machine learning

USC and Amazon establish Center for Secure and Trusted Machine Learning

January 28, 2021

Exterior of USC Viterbi School of Engineering

Center to support novel approaches to trust-centric machine learning and AI innovation.

Machine learning

ADePT: Auto-encoder based differentially private text transformation

Satyapriya Krishna, Rahul Gupta, Christophe Dupuy

EACL 2021

2021

Privacy is an important concern when building statistical models on data containing personal information. Differential privacy offers a strong definition of privacy and can be used to solve several privacy concerns (Dwork et al., 2014). Multiple solutions have been proposed for the differentially-private transformation of datasets containing sensitive information. However, such transformation algorithms

Machine learning

Amazon Scholars Michael Kearns and Aaron Roth discuss the ethics of machine learning

Staff writer

October 19, 2020

Michael Kearns and Aaron Roth seated at a table in front of a large chalk board.

Two of the world’s leading experts on algorithmic bias look back at the events of the past year and reflect on what we’ve learned, what we’re still grappling with, and how far we have to go.

Machine learning

Security, privacy, and abuse prevention

Creating a secure suite of hardware, software, and services with privacy, while giving you control over your information.

A differentially private text perturbation method using a regularized Mahalanobis metric

Zekun Xu, Abhinav Aggarwal, Oluwaseyi Feyisetan, Nathanael Teissier

EMNLP 2020 Workshop on PrivateNLP

2020

Balancing the privacy-utility tradeoff is a crucial requirement of many practical machine learning systems that deal with sensitive customer data. A popular approach for privacy-preserving text analysis is noise injection, in which text data is ﬁrst mapped into a continuous embedding space, perturbed by sampling a spherical noise from an appropriate distribution, and then projected back to the discrete

Security, privacy, and abuse prevention

Automatic discovery of privacy–utility pareto fronts

Brendan Avent, Javier González, Tom Diethe, Andrei Paleyes, Borja Balle

PETS 2020

2020

Differential privacy is a mathematical framework for privacy-preserving data analysis. Changing the hyperparameters of a differentially private algorithm allows one to trade off privacy and utility in a principled way. Quantifying this trade-off in advance is essential to decision-makers tasked with deciding how much privacy can be provided in a particular application while maintaining acceptable utility

Machine learning

Amazon SageMaker solution for privacy in natural language processing

Theodore Vasiloudis, Ehsan M. Kermani

2020

More and more text data are becoming available these days to train Natural Language Processing models such as sentiment analysis, predictive keyboards and question-answering chatbots. If companies that deploy such models use data provided by users, they have a responsibility to take steps to ensure their users' privacy. In this solution we demonstrate how one can use Differential Privacy to build accurate

Conversational AI

Preserving privacy in analyses of textual data

Tom Diethe, Oluwaseyi Feyisetan, Borja Balle, Thomas Drake

WSDM 2020

2020

Amazon prides itself on being the most customer-centric company on earth. That means maintaining the highest possible standards of both security and privacy when dealing with customer data. This month, at the ACM Web Search and Data Mining (WSDM) Conference, my colleagues and I will describe a way to protect privacy during large-scale analyses of textual data supplied by customers. Our method works by,

Conversational AI

Differential privacy

Work with us