Secure Systems Group (SSG) at Aalto University & University of Waterloo

dawn-dynamic-adversarial-watermarking-of-neural-networks

Watermarking against model extraction attacks in MLaaS. ACM MM 2021.

WAFFLE

WAFFLE: Watermarking in Federated Learning

distribution-attestation

Code for the paper "Attesting Distributional Properties of Training Data for Machine Learning" to appear in ESORICS 2024.

sel4-tee

BliMe

Blinded Memory, a new architecture for secure outsourced computation

ad3-action-distribution-divergence-detector

Real-time Adversarial Perturbations against Deep Reinforcement Learning Policies: Attacks and Defenses

conflicts-in-ml-protection-mechanisms

Study of conflicts occurring between various ML protection mechanisms. AAAI 2023.

sok-unintended-interactions

.github

authenticated-data-structures

BliMe-firesim

Falseclaims

GrOVe

GrOVe: Ownership Verification of Graph Neural Networks using Embeddings. IEEE S&P 2024.

sel4-tee-manifest

semalloc

ssg2021StyleTransferStudy

A user study about style transfer.

BliMe-block-inclusivecache-sifive

BliMe-chipyard

BliMe-gem5

BliMe-gem5-experiments

BliMe-linter

BliMe-model

BliMe-riscv-boom

BliMe-riscv-tests

BliMe-rocket-chip

BliMe-testchipip

FLARE

FLARE: Fingerprinting Deep Reinforcement Learning Agents using Universal Adversarial Masks

openfl

An open framework for Federated Learning.

sel4-tee-mapping

ssg-research.github.io

SSG Website