EARLY-CAREER RESEARCHERS IN MEDICAL APPLICATIONS @ CERN – SHORT TALKS

Tuesday 18 Nov 2025, 16:00 → 17:00 Europe/Zurich

40/S2-D01 - Salle Dirac (CERN)

 

Discover how technological advances for high-energy physics have become essential tools for modern medicine. 

This series of short talks is given by Early-Career Researchers working on projects related to diverse medical applications that arise from technology developed at CERN and in high-energy physics.

In this seminar 2 Researchers, Heloisa Barbosa Da Silva and Paolo Cacace, will share their work on the Cafein project.

Talk 1: Breast Cancer Risk Prediction: Exploring AI, Federated Learning, and Interpretability Approaches

Talk 2: Federated Learning and efficient learning for Edge Based Biomedical Signal Analysis

You can also watch the seminar on zoom (click on videoconference)

For information about the next Knowledge Transfer Seminar, please sign up to our e-group at http://cern.ch/go/F9cX

16:00 → 16:05 Introduction

16:05 → 16:20 Breast Cancer Risk Prediction: Exploring AI, Federated Learning, and Interpretability Approaches

Breast cancer is a major health concern. It is one of the leading causes of cancer related death among women and one of the most commonly diagnosed type of cancer. The risk factors that influence the likelihood of developing this disease are numerous, including modifiable factors such as lifestyle and nutrition. Currently, breast cancer screening tools are tailored to detect the disease rather than to predict it. However, overdiagnosis and overtreatment remains significant concerns. Knowing who is at high risk is extremely important in determining for whom breast cancer screening is most effective and who should be prioritized.
In this project, using The European Prospective Investigation into Cancer and Nutrition – a long-term study with more than half million participants, as the database, several machine learning algorithms, including federated learning, were used to test and to evaluate the most relevant variables that affects the model’s output. From the results insights were found regarding lifestyle and nutritional variables information, for example root vegetables consumption, sweets consumption, alcohol consumption in a specific period of life, that could be useful for new researchers and studies.

Convener: Heloisa Barbosa Da Silva (Universidade de Coimbra (PT))

Heloisa_Seminar copy.pdf

Heloisa_Seminar copy.pptx

16:20 → 16:35 Federated Learning and efficient learning for Edge Based Biomedical Signal Analysis

My contribution introduces a federated and efficient deep learning framework for biomedical signal analysis, with a focus on ECG signal processing. The approach leverages technologies originally developed at CERN for time-series anomaly detection in industrial systems, adapting them to the healthcare domain to ensure scalability, efficiency, and data privacy. A hybrid architecture combining convolutional neural networks and Transformer attention layers was used to capture both fine-grained waveform patterns and long-term temporal dependencies in multichannel physiological signals.
The framework has been successfully tested on edge devices, demonstrating the feasibility of real-time inference under resource constraints. To enable personalized healthcare while preserving patient data privacy, further development will see those model federated using CAFEIN®, a federated learning platform developed at CERN. The objective is to demonstrate how cross domain knowledge transfer and efficient AI can drive the development of next-generation biomedical solutions operating at the edge.

Convener: Paolo Cacace (Sapienza Universita e INFN, Roma I (IT))

Introduction KT Seminar CAFEIN November 2025 final.pptx

Seminar_Paolo_Cacace_18_Nov.pdf

Seminar_Paolo_Cacace_18_Nov.pptx