illustration of lock, four asterisks and a shield with a fingerprint on it on a green gradient background

Security, privacy, and cryptography

Leveraging Large Language Models for Automated Proof Synthesis in Rust

photos of PhD students Anunay Kulshrestha and Karan Newatia with a photo of Senior Cryptographer Josh Benaloh in the middle with a microphone icon on a blue and purple gradient background

Intern Insights: Dr. Josh Benaloh with Anunay Kulshrestha and Karan Newatia

Bayes Security (β*, higher better) of DP-SGD against membership and attribute inference on the Adult and Purchase datasets as a function of the number of training steps. The green dashed line shows the corresponding DP budget (ε). Membership and attribute inference curves decrease monotonically as the model is trained for longer and the accuracy and DP budget grow.

Closed-Form Bounds for DP-SGD against Record-level Inference

ProvCam: A Camera Module with Self-Contained TCB for Producing Verifiable Videos

Current selections

Sort by: Most recent

Clear selections

Search within these results

Content Types

Published Date

Video

Introducing Microsoft’s New Open Source Fuzzing Platform – CppCon

March 29, 2021 | Mike Walker

This native code security talk is a joint presentation by Principals from Windows Security (COSINE) and Microsoft Research. The work by Google and other contributors to the llvm ecosystem on libfuzzer, ASan, and sancov have…

51:46

Publication

Private Non-smooth Empirical Risk Minimization and Stochastic Convex Optimization in Subquadratic Steps

Janardhan (Jana) Kulkarni, Yin Tat Lee, Daogao Liu

NeurIPS 2021 | March 2021

Project

Publication

U.S. Broadband Coverage Data Set: A Differentially Private Data Release

Mayana Pereira, Allen Kim, Joshua Allen, Kevin White, Juan M. Lavista Ferres, Rahul Dodhia

March 2021

Project Project Project

Tool

Synthetic data showcase

Generates synthetic data and user interfaces for privacy-preserving data sharing and analysis. In many cases, the best way to share sensitive datasets is not to share the actual sensitive datasets, but user interfaces to derived…

GitHub

Tool

Twin Smooth Integers

This code was written for conducting experiments that are published in an academic paper at Eurocrypt to find twin smooth integers. The code does not contain any cryptographic algorithms, but can be used to find…

GitHub

Tool

EverParse

EverParse is a framework for automatically generating fully verified secure parsers from data format specifications described in domain-specific languages. It consists of a verified F* combinator library, and untrusted frontends compiling from data format specifications…

GitHub Publication

Publication

Enclosure: language-based restriction of untrusted libraries

Adrien Ghosn

ASPLOS | March 2021

Publication

Training Data Leakage Analysis in Language Models

Huseyin A. Inan, Osman Ramadan, Lukas Wutschitz, Daniel Jones, Victor Rühle, James Withers, Robert Sim

February 2021

Event

Microsoft at NDSS 2021

February 21, 2021 – February 24, 2021 | Virtual

Microsoft is proud to be a sponsor of the Network and Distributed System Security Symposium (NDSS) 2021. Stop by our virtual booth to learn more about our research and open opportunities at Microsoft.

Publication

Stronger NAS with Weaker Predictors

Junru Wu, Xiyang Dai, Dongdong Chen, Yinpeng Chen, Mengchen Liu, Ye Yu, Zhangyang Wang, Zicheng Liu, Mei Chen, Lu Yuan

NeurIPS 2021 | February 2021

Security, privacy, and cryptography

Highlights