Horário: 14h
Palestrante: Leia Sousa de Sousa (doutorado)
Orientador: Prof André Drummond
Título: Engenharia de tráfego em redes ópticas metropolitanas
Resumo: Com os crescentes requisitos de largura de banda para as pessoas e o desenvolvimento da urbanização, o movimento da população na cidade tem uma influência crescente na distribuição do tráfego nas dimensões de espaço e tempo. A distribuição desequilibrada resulta no bloqueio de serviços em diferentes áreas e em diferentes períodos de tempo e reduz a utilização de recursos de espectro em redes ópticas elásticas. Esta pesquisa propõe um algoritmo de alocação que foca na redução do bloqueio geral e por área.
Horário: 14h20
Palestrante: George Oliveira Barros (doutorado)
Orientador: Prof Flavio Vidal
Título: Métodos Computacionais para Identificação Automática de Podocitopatia em Imagens de Glomérulos Renais
Abstract: Podocyte lesions in renal glomeruli are identified by pathologists using visual analyses of kidney tissue sections (histological images). By applying automatic visual diagnosis systems, one may reduce the subjectivity of analyses, accelerate the diagnosis process, and improve medical decision accuracy. Towards this direction, we present here a new data set of renal glomeruli histological images for podocitopathy classification and a deep neural network model -- called here Pathospotter-PodNet. The data set consists of 835 digital images (374 with podocytopathy and 430 without podocytopathy), annotated by a group of pathologists. Our proposed classification method is grounded on a deep neural networks (DNN) (VGG19 pretrained) and was compared with other six state-of-the-art models in two dataset versions (RGB and gray level) and two different training contexts: pre-trained models (transfer learning) and from-scratch, both with hyper-parameters tuning. The PodNet method achieved classification results to 90.9\% of f1-score, 88.9\% precision, and 93.2\% of recall in the final validation sets. Our results suggest that computational approaches based on DNNs are promising tools for the medical diagnosis of podocitopathy.
Horário: 14h40
Palestrante: Liriam Michi Enamoto (doutorado)
Orientador: Prof Li Weigang
Título: Generic Multimodal Gradient-Based Meta Learning Framework
Resumo: Research in natural language processing, bio-medicine, and computer vision achieved excellent results in machine learning due to the success of the Transformer-based models. However, these excellent results depend on the labeled high-quality and large-scale datasets. If one of these requirements is not met, the model may lack generalization ability, and its performance will be unsatisfactory. To address these issues, this research proposes a Generic Multimodal Gradient-Based Meta Framework (GeMGF) trained from scratch to avoid language bias, learns from a few data, and reduces the model degradation trained on a finite dataset. GeMGF was evaluated using the benchmark dataset CUB-200-2011 for the text and image classification tasks. The results show that GeMGF is simple, efficient, and adaptable to other data modalities and fields.
Horário: 15h
Palestrante: Lucas Fernandes Ribeiro (doutorado)
Orientador: Prof Ricardo Jacobi
Title: Evaluating a Machine Learning-based Approach for Cache Configuration
Abstract: As the systems perform progressively complex tasks, the search for energy efficiency in computational systems is constantly increasing. The cache memory has a fundamental role in this issue. Through dynamic cache reconfiguration techniques, it is possible to obtain an optimal cache configuration that minimizes the impacts of energy losses. To achieve this goal, a precise selection of cache parameters plays a fundamental role. In this work, a machine learning-based approach is evaluated to predict the optimal cache configuration for different applications considering their dynamic instructions and a variety of cache parameters, followed by experiments showing that using a smaller set of application instructions it is already possible to obtain good classification results from the proposed model. The results show that the model obtains an accuracy of 96.19% using the complete set of RISC-V instructions and 96.33% accuracy using the memory instructions set, a more concise set of instructions that directly affect the cache power model, besides decreasing the model complexity.
Horário: 15h20
Palestrante: João Paulo Costa de Araújo (mestrado)
Orientadora: Profa Genaína Rodrigues
Title: Applying Artificial Immune Systems Principles to Cyber-Physical Systems
Abstract: Proposição de uma metodologia para aumentar a capacidade de monitorar e analisar com segurança sistemas ciber-físicos para contextos de execução com incerteza, aplicando princípios de sistemas imunológicos artificiais.
Horário: 15h40
Palestrante: Hércules de Campos Júnior (mestrado)
Orientadora: Profa Priscila Barreto
Title: 5G - ORAN
Abstract: a definir
Local: Teams MS - Equipe PPGI-316415 Seminário, Canal Seminários 1-2022
Profa Célia Ghedini Ralha (Este endereço de email está sendo protegido de spambots. Você precisa do JavaScript ativado para vê-lo.)
Coordenadora Seminários de Pós-Graduação em Informática 1-2022