Irippuge Milinda Perera
Irippuge Milinda Perera is a Software Engineer in the Security & Privacy group at Google. He received his Ph.D. in Computer Science from the City University of New York (CUNY) in 2015. His research interests are in data anonymization, mobile security, steganography, and cryptography.
Research Areas
Authored Publications
Sort By
Google COVID-19 Vaccination Search Insights: Anonymization Process Description
Gregory A. Wellenius
Evgeniy Gabrilovich
Chaitanya Kamath
Damien Desfontaines
Arti Patankar
Krishna Kumar Gadepalli
Akim Kumok
Adam Boulanger
Mark Young
Tague Griffith
Karen Lee Smith
Charlotte Stanton
Yael Mayer
Chris Scott
Shailesh Bavadekar
John S. Davis
Benjamin Miller
Arxiv.org (2021)
Preview abstract
This report describes the aggregation and anonymization process applied to the COVID-19 Vaccination Search Insights~\cite{vaccination}, a publicly available dataset showing aggregated and anonymized trends in Google searches related to COVID-19 vaccination. The applied anonymization techniques protect every user’s daily search activity related to COVID-19 vaccinations with $(\varepsilon, \delta)$-differential privacy for $\varepsilon = 2.19$ and $\delta = 10^{-5}$.
View details
A General Purpose Transpiler for Fully Homomorphic Encryption
Alain Forget
David Marn
Eric P. Astor
Rob Springer
Damien Desfontaines
Yurii Sushko
Bernat Guillen
Aishe Krishnamurthy
Xinyu Ye
Phillipp Schoppmann
Itai Zukerman
Rafael Misoczki
Shruthi Gorantala
Cameron Tew
Sasha Kulankhina
Sean Purser-Haskell
Sam Ruth
Miguel Guevara
Asra Ali
Google LLC (2021)
Preview abstract
Fully homomorphic encryption (FHE) is an encryption scheme which enables computation on encrypted data without revealing the underlying data. While there have been many advances in the field of FHE, developing programs using FHE still requires expertise in cryptography. In this white paper, we present a fully homomorphic encryption transpiler that allows developers to convert high-level code (e.g., C++) that works on unencrypted data into high-level code that operates on encrypted data. Thus, our transpiler makes transformations possible on encrypted data.
Our transpiler builds on Google's open-source XLS SDK (https://github.com/google/xls) and uses an off-the-shelf FHE library, TFHE (https://tfhe.github.io/tfhe/), to perform low-level FHE operations. The transpiler design is modular, which means the underlying FHE library as well as the high-level input and output languages can vary. This modularity will help accelerate FHE research by providing an easy way to compare arbitrary programs in different FHE schemes side-by-side. We hope this lays the groundwork for eventual easy adoption of FHE by software developers. As a proof-of-concept, we are releasing an experimental transpiler (https://github.com/google/fully-homomorphic-encryption/tree/main/transpiler) as open-source software.
View details
Google COVID-19 Search Trends Symptoms Dataset: Anonymization Process Description
Katie Everett
Gerardo Flores
Ilya Eckstein
Evgeniy Gabrilovich
Chaitanya Kamath
Damien Desfontaines
Krishna Kumar Gadepalli
Akim Kumok
Thomas Ludwig Roessler
Rayman Huang
Charlotte Stanton
Yael Mayer
Shailesh Bavadekar
John S. Davis
Venky Ramachandran
Gregory Alexander Wellenius
Arxiv.org, N/A (2020)
Preview abstract
This report describes the aggregation and anonymization process applied to the initial version of COVID-19 Search Trends symptoms dataset, a publicly available dataset that shows aggregated, anonymized trends in Google searches for symptoms (and some related topics). The anonymization process is designed to protect the daily search activity of every user with \varepsilon-differential privacy for \varepsilon = 1.68.
View details
KHyperLogLog: Estimating Reidentifiability and Joinability of Large Data at Scale
Wei-Yen Day
Damien Desfontaines
Chao Li
Daniel Simmons-Marengo
Pern Hui Chia
Miguel Guevara
Qiushi Wang
Proceedings of the 2019 IEEE Symposium on Security and Privacy
Preview abstract
Understanding the privacy relevant characteristics of data sets, such as reidentifiability and joinability, is crucial for data governance, yet can be difficult for large data sets. While computing the data characteristics by brute force is straightforward, the scale of systems and data collected by large organizations demands an efficient approach. We present KHyperLogLog (KHLL), an algorithm based on approximate counting techniques that can estimate the reidentifiability and joinability risks of very large databases using linear runtime and minimal memory. KHLL enables one to measure reidentifiability of data quantitatively, rather than based on expert judgement or manual reviews. Meanwhile, joinability analysis using KHLL helps ensure the separation of pseudonymous and identified data sets. We describe how organizations can use KHLL to improve protection of user privacy. The efficiency of KHLL allows one to schedule periodic analyses that detect any deviations from the expected risks over time as a regression test for privacy. We validate the performance and accuracy of KHLL through experiments using proprietary and publicly available data sets
View details
