PhD Candidate · Loughborough University

Petros Andreou

I study machine unlearning: methods for removing the influence of selected training data from AI models without training them again from scratch. The goal is to make that removal reliable and verifiable.

Conference presentation Presented at WIPE-OUT 2, ECML PKDD 2026, Naples, Italy - 7 September 2026

Google Scholar GitHub arXiv

Office: N.2.17, Haslegrave Building, Department of Computer Science, Loughborough University, Loughborough LE11 3TU, United Kingdom

CONCEPTUAL OVERVIEW Evaluating unlearning

Trained AI model

Apply an unlearning method

Model to evaluate

Two questions

Evidence of removal What do the tests show about the influence of the data requested for removal?

Useful ability retained How well does the model still perform on data it should retain?

A conceptual flow, not an experimental result.

Publications

arXiv Accepted 2606.00380 1 Jun 2026

SUPREME: A Multi-GPU Framework for Reproducible Image Unlearning Method Evaluation

Petros Andreou , Jamie Lanyon , Axel Finke , Georgina Cosma

Accepted at WIPE-OUT 2: 2nd Workshop on Machine Unlearning and Privacy Preservation, ECML PKDD 2026, Naples, Italy

Read on arXiv View code Read on LinkedIn
arXiv Accepted 2510.26714 30 Oct 2025

On the importance of multiple training seeds for evaluating machine unlearning

Jamie Lanyon , Axel Finke , Petros Andreou , Georgina Cosma

Accepted at WIPE-OUT 2: 2nd Workshop on Machine Unlearning and Privacy Preservation, ECML PKDD 2026, Naples, Italy

Read on arXiv