Computer Science » Department Events

03/11 2pm-3pm: Systems and Security Seminar - From End-to-end verifiable Voting Systems to Fully Verified Elections (Peter Ryan)

Computer Science — Mon, 19 Oct 2026 14:19:46 GMT

When: 14:00 - 15:00, Tue, 03 Nov '26

Where: CS101

Title: End-to-end verifiable Voting Systems to Fully Verified Elections

By Peter Y A Ryan

Abstract:

The idea of end-to-end verifiable (E2E V) voting schemes has been around for over three decades. The goal is to ensure that any manipulation of any vote from the point of casting to the point of tallying is detectable by voters and/or observers, and ideally correctable. This should be achieved with minimal trust assumptions without trusting the correctness of the code running on voting machines, servers etc. At the same time, we need to guarantee conflicting properties like vote, privacy, coercion resistance etc.

E2E V is a laudable goal, essentially election dependability for the people by the people, and many excellent schemes have been proposed, and some prototyped and trialed, but to date none have seen widespread, long-term deployment. The begs the question: why has there been so little uptake of such schemes even though they are manifestly more trustworthy that the commercial systems that are in use? We suggest that a large part of the reluctance to deploy E2E V systems the need for voters to perform certain checks on encrypted ballots, the so-called “individual” checks. Such checks are a problem for usability, understandability and acceptability. Furthermore, if too few voters perform their checks, then the election cannot be deemed to have been verified, even if run on a verifiable system.

I will present an in-person scheme in which these checks are outsourced to independent auditors and for voters the experience is a simple and familiar vote and go. There is no longer any need or expectation for voters perform any checks (but the option can still be made available to them). In essence the idea is for the Election Authority to commit in advance to ballots that carry encryptions of the voting options. These can be audited before, during and after voting by independent observers and auditors. More exactly, what is committed to in the setup phase is ballots with malleable signatures over trivial encryptions of the voting options, allowing the options to be presented to the voters in plaintext. Ballot auditing now reduces to signature verification, which can be performed by anyone and is dispute-free. We also propose techniques to outsource the recorded-as-cast checks, that are also usually regarded as individual.

Thus, voters simply present their vote selection in plaintext to a scanner in the booth and then present the resulting printout showing a re-encryption of the vote along with transformed signature to the election officials. Assuming that independent auditors and observers perform sufficient audits, we have elections that are in fact verified rather than just verifiable. We refer to schemes with this property as “fully verified”. Time permitting, I will touch on the challenges of achieving this goal for online voting.

Short bio:

Peter Ryan is full Professor of Applied Security at the University of Luxembourg since Feb 2009, Emeritus from July 2025. He has over 30 years of experience in cryptography, information assurance and formal verification. He pioneered the application of process calculi to modelling and analysis of secure systems, in particular the characterization of non-interference and the analysis of security protocols. He has published extensively on cryptography, cryptographic protocols, security policies, mathematical models of computer security and, more recently, voter-verifiable election systems, Quantum Authenticated Key Establishment Protocols and post-quantum crypto.

He was awarded a PhD in mathematical physics from the University of London in 1982. He sits on or has sat on the program committees of numerous, prestigious security conferences, notably: IEEE Security and Privacy, IEEE Computer Security Foundations Workshop/Symposium (CSF), the European Symposium on Research in Computer Security (ESORICS). From 1999 to 2007 he was the President of the ESORICS Steering Committee. In 2016 he founded the Verifiable Voting series of workshops associated with Financial Crypto. He is a Visiting Professor at Surrey University and the ENS Paris.

07/12 2pm-3pm: TIA Centre Seminar Series - Shubham Innani, Indiana University School of Medicine, USA

Computer Science — Wed, 14 Oct 2026 16:55:46 GMT

When: 14:00 - 15:00, Mon, 07 Dec '26

Explainable Artificial Intelligence (XAI) in Computational Pathology.

Computational pathology models are increasingly used for diagnosis, prognosis, and treatment prediction, yet many remain difficult to interpret. Drawing on a recent MICCAI SIG-CompPath review, this talk presents a pathology-focused framework for explainable AI, including a structured taxonomy of XAI methods and practical guidance for linking clinical questions to appropriate interpretability approaches.

30/11 2pm-3pm: TIA Centre Seminar Series - Dr Frédérique Meeuwsen, Radboud UMC, The Netherlands

Computer Science — Wed, 14 Oct 2026 16:54:20 GMT

When: 14:00 - 15:00, Mon, 30 Nov '26

16/11 2pm-3pm: TIA Centre Seminar Series - Dr Hao Chen, The Hong Kong University of Science and Technology, China

Computer Science — Wed, 14 Oct 2026 16:52:51 GMT

When: 14:00 - 15:00, Mon, 16 Nov '26

Pathology Foundation Models-driven Agentic AI for Precision Oncology.
With the deep integration of AI and digital pathology, large pathology models become the foundation backbones in computational pathology for precise cancer diagnosis and treatment. This talk will systematically explore the cutting-edge advancements of large pathological models in precise cancer diagnosis, molecular subtyping prediction, treatment response assessment, and prognostic analysis, revealing the breakthrough potential in achieving "micro-meso-macro" analysis through multimodal data integration (such as whole-slice images, imaging, genomics, and clinical information). We will further explore how large pathology models-driven agentic AI are transforming the paradigm in precision oncology and beyond.

09/11 2pm-3pm: TIA Centre Seminar Series - Dr Jan Clusmann, TU Dresden, Germany

Computer Science — Wed, 14 Oct 2026 16:49:52 GMT

When: 14:00 - 15:00, Mon, 09 Nov '26

Safety and Security of Large Language Models in Healthcare.
Large language models (LLMs) are rapidly entering clinical practice. Drawing on insights from a recent Nature Review, Germany’s largest survey of clinician AI chatbot use (>4,000 respondents), and real-world deployment experience, Dr Jan Clusmann explores the opportunities, risks, and readiness of LLMs for healthcare.

09/11 2pm-3pm: Computer Science Education Research Group Seminars

Computer Science — Wed, 30 Sep 2026 14:03:06 GMT

When: 14:00 - 15:00, Mon, 09 Nov '26

Where: CS0.05

Join the CS Education Research Group for our regular term-time seminars! These sessions provide a platform to discuss and share ongoing work from both internal and external researchers.

When: Weeks 2, 4, 6 & 8 | 2:00 PM – 3:00 PM
Where: Room CS0.05

23/11 2pm-3pm: Computer Science Education Research Group Seminars

Computer Science — Wed, 30 Sep 2026 14:03:06 GMT

When: 14:00 - 15:00, Mon, 23 Nov '26

Where: CS0.05

Join the CS Education Research Group for our regular term-time seminars! These sessions provide a platform to discuss and share ongoing work from both internal and external researchers.

When: Weeks 2, 4, 6 & 8 | 2:00 PM – 3:00 PM
Where: Room CS0.05