Seminars

  • Seminars

    The Department of Computer Science invites everyone to the lectures of the PPGI Seminar discipline, with the themes ( From log4j to mapping vulnerable maven dependencies in a Brazilian institution / Automatic Lifting of Static Taint Analyses for Evolving Product Lines)

  • Seminars

    The Department of Computer Science invites everyone to the lectures of the PPGI Seminar course, with the themes (A New Fine Tuning Approach to Improve the Interlinguistic Performance of LLMs / Towards a Theory for Source Code Rejuvenation / Quad-Tree Block-Based Variational Autoencoder for Point-Clouds / Evaluation of Named Entity Recognition Using Ensembles in Transformer Models for Brazilian Public Texts).

  • Seminars

    The Department of Computer Science invites everyone to the lectures of the PPGI Seminar course, with the themes (Development of an integrated attack detection and mitigation solution for software-defined networks / Unrestricted Approximate String Matching for Synthetic and DNA Sequences: Practical Analysis / Performance Evaluation of Large Language Models in Software Rejuvenation: Developing a Comparative Methodology).

  • Seminars

    The Department of Computer Science invites everyone to the lectures of the PPGI Seminar course, with the themes (Methodology for Producing Actionable Threat Intelligence at the Tactical Level / Comparative Analysis of Data Protection Laws and Privacy Frameworks: Optimizing Solutions for LGPD Compliance)

+

Seminários - 26/05/17 - 14h

CUDA-Sankoff: using GPU to accelerate the pairwise structural RNA alignment

/Transport Mode Detection Using Mobile Phones and Artificial Neural Networks for Accident Prevention

/Estudo de Caso de uma Aplicação de Linha de Produtos a um Projeto de Dotação Militar

Local: Sala Multiuso CIC

Horário: 14h

Palestrante: Daniel Sundfeld Lima (doutorado)

Título: CUDA-Sankoff: using GPU to accelerate the pairwise structural RNA alignment

Resumo: In this work, we propose and evaluate CUDA-Sankoff, a solution to the RNA structural alignment problem based on the Sankoff algorithm in Graphics Processing Units (GPUs). To our knowledge, this is the first time the Sankoff algorithm is implemented in GPU. In our solution, we show how to linearize the Sankoff 4-dimensional dynamic programming (4D DP) matrix and we propose a two-level wavefront approach to exploit the parallelism. The results were obtained with two different NVidia GPUs, comparing sets of real RNA sequences with  lengths from 46 to 281 nucleotides. We show that our GPU approach is up to 24 times faster than a 16-core CPU solution in the 281 nucleotide Sankoff execution.

 

Horário: 14h30

Palestrante: Mateus Mendelson da Silva (mestrado)

Título: Transport Mode Detection Using Mobile Phones and Artificial Neural Networks for Accident Prevention

Resumo: Este artigo apresenta um método baseado em Redes Neurais Artificiais que, utilizando sinais capturados a partir de acelerômetros de telefones celulares convencionais, é capaz de identificar determinados padrões de deslocamento do usuário definidos pelo local onde tal dispositivo é armazenado (no bolso, na mão ou no console) e pela forma como o usuário se desloca (por meio de um carro ou a pé). Tal método possibilita a identificação de situações de risco de acidentes, como dirigir e andar utilizando um smartphone. Os testes mostram que o método proposto é capaz identificar uma entre 10 possíveis classes de transporte com uma taxa de acerto médio de 95,3%.

 

Horário: 15h

Palestrante: Ricardo de Lima (mestrado)

Título: Estudo de Caso de uma Aplicação de Linha de Produtos a um Projeto de Dotação Militar

 

Organizadora: Profa Célia Ghedini Ralha (This email address is being protected from spambots. You need JavaScript enabled to view it.)

Coordenadora dos Seminários de Pós-Graduação Informática 2017-1