Security, privacy, and cryptography
Video
CROSS — Leveraging AI ASICs for Homomorphic Encryption
Artificial Intelligence (AI) is driving a new industrial revolution, transforming human workflows increasingly into digital tokens, i.e., tokenizing the entire world. However, this transformation exposes sensitive data at an unprecedented scale, leading to heavy privacy…
Publication
Contextualized Privacy Defense for LLM Agents
Video
Hardware Realization and Implementation Security Evaluation of HQC, A NIST PQC Standard
Quantum computing is no longer a distant dream, its rapid progress is poised to revolutionize various fields from drug discovery to optimization. But this leap forward comes with a critical caveat: the pre-quantum public-key cryptographic…
Microsoft Research Blog
Media Authenticity Methods in Practice: Capabilities, Limitations, and Directions
As synthetic media grows, verifying what’s real, and the origin of content, matters more than ever. Our latest report explores media integrity and authentication methods, their limits, and practical paths toward trustworthy provenance across images,…
Video
Efficient Homomorphic Integer Computer from CKKS
Fully homomorphic encryption (FHE) has evolved from Gentry’s original blueprint into a diverse family of practical schemes, including BGV/BFV for exact arithmetic, DM/CGGI-style schemes for fast binary computation, and CKKS for high-throughput approximate arithmetic. I…
Project
FineACL
This project studies the data protection challenges when models and agents interact with sensitive data across multiple users. Large language models (LLMs) are increasingly deployed in enterprise settings where thLarge language models (LLMs) are increasingly…