Mariana Raykova
YOU?
Author Swipe
View article: On the Differential Privacy and Interactivity of Privacy Sandbox Reports
On the Differential Privacy and Interactivity of Privacy Sandbox Reports Open
The Privacy Sandbox initiative from Google includes APIs for enabling privacy-preserving advertising functionalities as part of the effort to limit third-party cookies. In particular, the Private Aggregation API (PAA) and the Attribution R…
View article: Provably Robust Watermarks for Open-Source Language Models
Provably Robust Watermarks for Open-Source Language Models Open
The recent explosion of high-quality language models has necessitated new methods for identifying AI-generated text. Watermarking is a leading solution and could prove to be an essential tool in the age of generative AI. Existing approache…
View article: Differentially Private Ad Conversion Measurement
Differentially Private Ad Conversion Measurement Open
In this work, we study ad conversion measurement, a central functionality in digital advertising, where an advertiser seeks to estimate advertiser website (or mobile app) conversions attributed to ad impressions that users have interacted …
View article: Differentially Private Ad Conversion Measurement
Differentially Private Ad Conversion Measurement Open
In this work, we study ad conversion measurement, a central functionality in digital advertising, where an advertiser seeks to estimate advertiser website (or mobile app) conversions attributed to ad impressions that users have interacted …
View article: UN Handbook on Privacy-Preserving Computation Techniques
UN Handbook on Privacy-Preserving Computation Techniques Open
This paper describes privacy-preserving approaches for the statistical analysis. It describes motivations for privacy-preserving approaches for the statistical analysis of sensitive data, presents examples of use cases where such methods m…
View article: Distributed, Private, Sparse Histograms in the Two-Server Model
Distributed, Private, Sparse Histograms in the Two-Server Model Open
We consider the computation of sparse, (ε, ϑ)-differentially private~(DP) histograms in the two-server model of secure multi-party computation~(MPC), which has recently gained traction in the context of privacy-preserving measurements of a…
View article: Advances and Open Problems in Federated Learning
Advances and Open Problems in Federated Learning Open
Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the trai…
View article: Think Global, Act Local: Gossip and Client Audits in Verifiable Data Structures
Think Global, Act Local: Gossip and Client Audits in Verifiable Data Structures Open
In recent years, there has been increasing recognition of the benefits of having services provide auditable logs of data, as demonstrated by the deployment of Certificate Transparency and the development of other transparency projects. Mos…
View article: Secure Single-Server Aggregation with (Poly)Logarithmic Overhead
Secure Single-Server Aggregation with (Poly)Logarithmic Overhead Open
Secure aggregation is a cryptographic primitive that enables a server to learn the sum of the vector inputs of many clients. Bonawitz et al. (CCS 2017) presented a construction that incurs computation and communication for each client line…
View article: PPML '19
PPML '19 Open
The area of privacy preserving machine learning has been of growing importance in practice, which has lead to an increased interest in this topic in both academia and industry. We have witnessed this through numerous papers and systems pub…
View article: Make Some ROOM for the Zeros
Make Some ROOM for the Zeros Open
Exploiting data sparsity is crucial for the scalability of many data analysis tasks. However, while there is an increasing interest in efficient secure computation protocols for distributed machine learning, data sparsity has so far not be…
View article: Secure Computation for Machine Learning With SPDZ
Secure Computation for Machine Learning With SPDZ Open
Secure Multi-Party Computation (MPC) is an area of cryptography that enables computation on sensitive data from multiple sources while maintaining privacy guarantees. However, theoretical MPC protocols often do not scale efficiently to rea…
View article: RapidChain
RapidChain Open
A major approach to overcoming the performance and scalability limitations of current blockchain protocols is to use sharding which is to split the overheads of processing transactions among multiple, smaller groups of nodes. These groups …
View article: Language Technologies in Teaching Bulgarian at Primary and Secondary School Level: The NBU Platform for Language Teaching (PLT)
Language Technologies in Teaching Bulgarian at Primary and Secondary School Level: The NBU Platform for Language Teaching (PLT) Open
The NBU Language Teaching Platform (PLT) is a versatile tool supporting language learning.So far used for FLSP etuition or blended learning in general foreign language classes, it is now being extended to provide e-support for teaching Bul…
View article: Privacy-Preserving Distributed Linear Regression on High-Dimensional Data
Privacy-Preserving Distributed Linear Regression on High-Dimensional Data Open
We propose privacy-preserving protocols for computing linear regression models, in the setting where the training dataset is vertically distributed among several parties. Our main contribution is a hybrid multi-party computation protocol t…
View article: 5Gen
5Gen Open
Secure multilinear maps (mmaps) have been shown to have remarkable applications in cryptography, such as multi-input functional encryption (MIFE) and program obfuscation. To date, there has been little evaluation of the performance of thes…