Advanced search
Start date
Betweenand

Privacy-preserving and backdoors defending: towards federated learning in medical settings

Grant number: 23/14759-0
Support Opportunities:Scholarships abroad - Research Internship - Doctorate
Effective date (Start): April 15, 2024
Effective date (End): April 14, 2025
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
Principal Investigator:Agma Juci Machado Traina
Grantee:Erikson Júlio de Aguiar
Supervisor: Abdelsalam Helal
Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
Research place: University of Florida, Gainesville (UF), United States  
Associated to the scholarship:21/08982-3 - Security and privacy in machine learning models to medical images against adversarial attacks, BP.DR

Abstract

In medical settings, data can be collected from heterogeneous sources, such as X-ray machines, blood exams, wearables, and Internet of Things devices. Artificial Intelligence (AI) strategies, such as Deep Learning (DL), can combine these data to discover patterns and anomalies. DL is specialized in classifying medical images to assist physicians in diagnosis. Even though AI can enhance the capacity of computational systems to solve complex problems, it can present security and privacy issues. Privacy concerns have gained attention in the last few years due to privacy regulations and security awareness against backdoor attacks that corrupt models. Federated Learning (FL) is a strategy that raises the privacy of the DL models because they train the models without sharing the sensitive dataset between the clients. However, FL can suffer from backdoor attacks and most robust privacy attacks, such as Membership Interference Attacks. We intend to develop new techniques to improve the privacy of the systems and strategies to defend against backdoors. The internship will be conducted at the University of Florida (UF), USA, supervised by Prof. Dr. Sumi Helal. The objectives of this project cover the interests of both the FAPESP thematic project ``Mining, Index and Visualizing Big Data in Clinical Decision Support Systems" -- (MIVisBD) and research carried out at UF by Prof. Helal.

News published in Agência FAPESP Newsletter about the scholarship:
Articles published in other media outlets (0 total):
More itemsLess items
VEICULO: TITULO (DATA)
VEICULO: TITULO (DATA)

Please report errors in scientific publications list using this form.