Banca de DEFESA: GABRIEL ARAUJO DE SOUZA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : GABRIEL ARAUJO DE SOUZA
DATE: 20/12/2022
TIME: 09:00
LOCAL: Sala do Google Meet - meet.google.com/qqy-keux-mys
TITLE:

Using Federated Learning Techniques to Improve Artificial Intelligence Models in the Context of Brazilian Public Institutions

KEY WORDS:

Federated Learning, Artificial Intelligence, Brazilian Public Institutions, Convergence Algorithms

PAGES: 144
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Sistemas de Computação
SPECIALTY: Arquitetura de Sistemas de Computação
SUMMARY:

The use of artificial intelligence models has become frequent in several areas of knowledge to resolve different problems efficiently. Due to this, many Brazilian Public Institutions have invested in AI solutions to improve and optimize their services. However, these institutions, mainly public safety organizations, use sensitive privacy data in their solutions. Thus, the use of this data is bureaucratic, primarily to respect all General Data Protection Law requirements. Furthermore, each institution explores a limited examples scenario which makes the AI models biased. The data sharing between institutions could provide the creation of general datasets with a better capacity to create more robust models. However, due to the nature of the data, this type of action is, in many cases, unfeasible. Thus, federated learning has gained space in the recent literature to enable the sharing of AI models safely. In this technique, instead of sharing data, only the models already trained are aggregated on a server to provide a new model. With this, it is possible to transfer knowledge from various models to create an improved version of them. Therefore, this work proposes using federated learning to create a safe environment for sharing AI models among Brazilian Public Institutions. In addition, the work proposes the experiment with different techniques present in the literature to identify the best federated algorithm used in this studied scenario.

COMMITTEE MEMBERS:
Presidente - 1678918 - NELIO ALESSANDRO AZEVEDO CACHO
Interno - 2316877 - EVERTON RANIELLY DE SOUSA CAVALCANTE
Externo ao Programa - 2579664 - ALLAN DE MEDEIROS MARTINS - UFRNExterno ao Programa - 1669545 - DANIEL SABINO AMORIM DE ARAUJO - UFRNExterno à Instituição - PEDRO MESQUITA MOURA