"Systems Security Research at KTH" workshop

 

 

In January 2026, Prof. N. Asokan was appointed as a part-time Visiting Wallenberg Chair at KTH. His mission is to further strengthen and expand systems security research at KTH by forging successful collaborations in systems security research with groups at KTH as well as in Swedish industry. Prof. Asokan’s research group, Secure Systems Group, will establish a presence at KTH in close collaboration with Network Systems Security Group, led by Prof. Panos Papadimitratos. The recruitment of two post-doctoral researchers at KTH, to be co-supervised by Professors Asokan and Papadimitratos, is already under way. Two doctoral students will be recruited later this year.

 

On Friday, April 24, afternoon, SSG and NSS will jointly organize a workshop to introduce SSG and the types of research themes and collaborations it is developing at KTH. The speakers include former graduates of SSG as well as experts in Sweden with whom SSG is building collaborations. The talks will span the two major themes that underlie systems security research at SSG: (1) Machine Learning (ML) and security/privacy: the interplay between ML and cybersecurity/privacy, and (2) Platform security: hardware-assisted software protection.

 

There is limited seating at the venue, but we also hope to broadcast the talks via Zoom.

 

Speakers

 

The workshop will be hybrid - in person attendance at the Herik Eriksson room, at Lindstedtsvägen 3 on the main KTH campus and online over zoom.

 

Please register using this KTH form by April 4, 2026.

 

More information on speakers and talks:

 

Prof. N. Asokan

Title: A brief introduction to Secure Systems Group (SSG)

Abstract: I will briefly describe recent research in SSG along both of SSG’s research themes (“platform security” and “ML and security/privacy”), and introduce the speakers.

Bio: Asokan was recently named to a Wallenberg Chair as a visiting professor at KTH. He is a professor of computer science at the University of Waterloo where he held a Cheriton Chair (2019–2025) and served as the Executive Director of the Cybersecurity and Privacy Institute. Asokan has been working in the broad area of systems security for over three decades, first in industry research labs and then in academia in Europe and Canada. He is a fellow of the ACM, the IEEE, and the Royal Society of Canada.

 

Prof. Sebastian Szyller

Title: Pinky Promise or Ironclad: Guarantees in Trustworthy Machine Learning

Abstract: Cryptography is built on a mathematical bedrock -- security guarantees that hold under well-defined assumptions that break only when violated. Similarly, differential privacy shows what a rigorous guarantee looks like: a formal, quantifiable bound on privacy loss that holds adversarially and composes predictably. Without differential privacy, many common privacy-preserving techniques offer no worst-case guarantees. Their protection depends on empirical evaluations that are necessarily incomplete. Through this lens, I will discuss where trustworthy ML stands today -- why the current empirical paradigm falls short for regulators, policymakers, and engineers deploying machine learning in high-risk settings. I'll discuss what a path towards stronger guarantees might look like, and the work being done to get there.

Bio: Sebastian Szyller is a recently appointed assistant professor at Aalto University where he leads the Trustworthy & Adversarial Computing Lab. He obtained his PhD from Aalto University in 2023, which was recognised with the Aalto Distinguished Dissertation Award and the Finnish AI Society Best Dissertation Award. His research sits at the intersection of machine learning, and security and privacy -- particularly model provenance and the formal foundations of trustworthy machine learning, with a focus on quantifiable guarantees and worst-case analysis. Prior to his current role, Sebastian was a research scientist at Intel Labs, where his work on provenance contributed to the C2PA standard. Across academia and industry, he has been working on trustworthy machine learning for ten years.

 

Prof. Buse Atli

Title: Threat Modelling for AI-Based Radio Access Networks (AI-RAN): Bridging the Gap Between Theory and Practice

Abstract: Threat modeling is a key security technique in both academia and industry for identifying and categorizing risks, vulnerabilities, and potential attacks on a system. It is widely used in AI/ML security as well. However, when these academic models are applied to real systems, practical limitations quickly emerge. Many theoretical attacks assume unconstrained adversaries who can freely manipulate inputs or send unlimited queries to AI/ML models, yet these assumptions often fail in real-world environments. Protocol, physical, and operational constraints can make some attacks infeasible, while proposed defenses may be hard to scale or incompatible with performance and reliability needs. In this talk, I examine how academic threat models translate into practice in AI-based Radio Access Network (AI-RAN) applications, outline what is realistically achievable in mobile networks, and argue for threat modeling that considers the full system and application context, not just isolated AI/ML components.

Bio: Buse Atli is a recently appointed Assistant Professor in the Cybersecurity division at Linköping University (LiU) and WASP AI/MLX Fellow. Before joining LiU, she served as a security researcher at Nokia Bell Labs, where she specialized in the development of threat modeling strategies designed to address the security and privacy challenges associated with the integration of Artificial Intelligence (AI) into network systems. She obtained her PhD degree from Aalto University in Finland. Her current research focuses on trustworthy machine learning, covering key areas such as verifiable protection mechanisms for machine learning systems, data/model provenance, and AI governance.

 

Dr. Thomas Nyman, and Dr. Merve Gülmez

Title: Memory Safety in Telecommunications with CHERI?

Abstract: The C and C++ programming languages are consistently the preferred languages for systems programming, embedded systems, and performance-critical applications, including many performance-critical components in telecommunications systems. In recent years, leading cybersecurity organization such as the US CISA, NSA, and other national cybersecurity agencies have shown an increased interest in systematically addressing memory-safety issues commonly associated with C and C++, urging software manufacturers to adopt memory-safe system programming languages, such as Rust, in favor of C and C++.

CHERI, which stands for Capability Hardware Enhanced RISC Instructions, is a hardware research architecture developed by the University of Cambridge and SRI International which presents an alternative to rewriting existing C and C++ code into another language. CHERI changes how software accesses memory, with the aim of systematically removing memory-safety vulnerabilities in C and C++.

In this talk, Merve Gülmez and Thomas Nyman will give an overview of Ericsson Research's assessment of CHERI's applicability to telecommunications systems, discuss some of the open challenges for the adoption of CHERI, and the joint research with SSG and other academic partners to enhance CHERI towards increased memory-safety coverage and accuracy, new threat models, and better performance.

Bios: Merve Gülmez is a researcher at Ericsson Security Research. She holds a PhD in Computer Science from KU Leuven, and her research focuses on systems security, memory safety, and improving software resilience.

Thomas Nyman works in Ericsson Product Security as an Expert in Trusted Hardware & Software Technologies. He holds a PhD in Computer Science from Aalto University, Finland, where his PhD dissertation was on the topic of hardware-assisted run-time protection in mobile and embedded systems. He leads the C/C++ Compiler Hardening Guide initiative under the Open Source Security Foundation (OpenSSF) Best Practices Working Group.

 

Prof. Panos Papadimitratos

Title: A brief introduction to the Networked Systems Security (NSS) group

Abstract: The talk will provide an overview of recent and ongoing NSS activities, also connecting to the SSG themes.

Bio: Panos Papadimitratos (Fellow, IEEE) earned his Ph.D. degree from Cornell University, Ithaca, NY. His research agenda includes a gamut of security and privacy problems, with emphasis on wireless and mobile networks. He chairs the ACM WiSec conference steering committee. Panos is a Knut and Alice Wallenberg Academy Fellow, a Fellow of the Young Academy of Europe, and an ACM Distinguished Member.

 

 

 

Back to the top