The Secure, Reliable, and Intelligent Systems Lab (SRI) is a research group in the Department of Computer Science at ETH Zurich. Our current research focus is on the areas of reliable, secure, robust and fair machine learning, probabilistic and quantum programming, and machine learning for code. Our work led to three ETH spin-offs: (AI for Code), ChainSecurity (security verification), and LatticeFlow (robust machine learning). Please see Research and Publications to learn more.

Latest Blog Posts

Latest News

Latest News & Blog Posts

LAMP: Extracting Text from Gradients with Language Model Priors: In this work we present an attack on federated learning's privacy specific to the text domain. We show that federated learning in the text domain can expose a lot of user data.

Timon Gehr, former doctoral student and current postdoctoral researcher at SRI Lab, has won the ETH Medal for his outstanding doctoral thesis. See website of D-INFK.

SRI Lab at ICLR 2022: SRI Lab will present five works at ICLR 2022! In this meta post we aggregate all content related to our ICLR papers, including links to the conference portal and individual blogposts where you can learn more about the topics we currently focus on.

7-8 October 2022: Workshop on Dependable and Secure Software Systems, hosting leading scientists who will present the latest research and most advanced methods for addressing this fundamental challenge. Website

Generating provably robust adversarial examples: We introduce the concept of provably robust adversarial examples. These are adversarial examples that are generated together with a region around them that can be proven robust to perturbations. We also show a method for generating large such regions in a scalable manner.

Professor Martin Vechev was appointed Full Professor of Computer Science in the Department of Computer Science. His achievements in a number of areas are globally regarded as groundbreaking.

Multi-neuron relaxation guided branch-and-bound: Learn more about how multi-neuron constraints can be used in a Branch-and-Bound framework to build a state-of-the-art complete neural network verifier.

Our new ETH spin-off LatticeFlow raises USD 2.8M to help companies build and deploy trustworty AI. Read articles on TechCrunch and ETH.