3IA PhD/Postdoc Seminar #44

Flash presentation: MRI-Ultrasound Image Registration for Prostate Cancer Care

Abstract: 3D registration between magnetic resonance imaging (MRI) and transrectal ultrasound (TRUS) plays a crucial role in the management of prostate cancer, enabling precise targeting of the lesions during biopsy planning and radiotherapy. MRI provides superior resolution and better lesion visualization, while TRUS offers real-time guidance during interventions. Combining the complementary properties of these two imaging modalities with artificial intelligence solutions helps improve the accuracy of biopsy and radiotherapy procedures.

Attribute Inference Attacks for Federated Regression Tasks

Abstract: Federated Learning (FL) enables multiple clients, such as mobile phones and IoT devices, to collaboratively train a global machine learning model while keeping their data localized. However, recent studies have revealed that the training phase of FL is vulnerable to reconstruction attacks, such as attribute inference attacks (AIA), where adversaries exploit exchanged messages and auxiliary public information to uncover sensitive attributes of targeted clients. While these attacks have been extensively studied in the context of classification tasks, their impact on regression tasks remains largely unexplored. In this paper, we address this gap by proposing novel model-based AIAs specifically designed for regression tasks in FL environments. Our approach considers scenarios where adversaries can either eavesdrop on exchanged messages or directly interfere with the training process. We benchmark our proposed attacks against state-of-the-art methods using real-world datasets. The results demonstrate a significant increase in reconstruction accuracy, particularly in heterogeneous client datasets, a common scenario in FL. The efficacy of our model-based AIAs makes them better candidates for empirically quantifying privacy leakage for federated regression tasks. 

Asteroids in the Digital Age: Insights from Modern Sky Surveys

Abstract: Asteroids are relics of planet formation, leftover building blocks that never became complete planets. Their importance extends far beyond astronomy: these space rocks record the history of the formation of our solar system, contain valuable resources such as metals and water, and some even pose potential threats to Earth. The study of asteroids is therefore of great scientific and practical importance. Until recently, we had detailed information about only a small number of asteroids. While more than a million minor planets have been discovered, only a few thousand were well studied before the age of digital sky surveys. The introduction of large-scale digital surveys has dramatically changed our understanding. Modern surveys use powerful telescopes and advanced detection technology to automatically scan the night sky, producing vast amounts of new data. This talk will discuss how projects such as the Sloan Digital Sky Survey (SDSS), SkyMapper, and Gaia have greatly improved our knowledge of asteroids. What challenges have we faced in analyzing the data? It will also look ahead to upcoming projects that promise even more discoveries.

Open discussion about the two contributions