Raul Sena Ferreira

Raul Sena Ferreira
Paul Sabatier University - Toulouse III | UPS Toulouse · Laboratoire d'Analyse et d'Architecture des Systèmes - UPR 8001 - LAAS

PhD Student at LAAS - CNRS
ML in dynamic environments, computer vision, and safe AI

About

26
Publications
3,900
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
45
Citations
Introduction
Interest in machine learning, statistical learning, and Safe AI for autonomous vehicles. (www.raulferreira.com.br)
Additional affiliations
September 2016 - September 2018
Federal University of Rio de Janeiro
Position
  • IT Coordinator at Instituto de Pesquisa Econômica e Aplicada (IPEA)
October 2015 - present
National Museum of Rio de Janeiro
Position
  • IT Specialist
March 2010 - December 2014
Federal Rural University of Rio de Janeiro
Position
  • Student

Publications

Publications (26)
Conference Paper
Full-text available
According to estimates from the Federal Government, 200,000 people disappear from their homes every year in Brazil. In 2002, Brazilian national network for identification and location of missing children and adolescents established four important goals to tackle this serious problem. Our work is motivated by two of them: to create an unified regist...
Conference Paper
Full-text available
Machine learning solutions usually consider that the train and test data has the same probabilistic distribution, that is, the data is stationary. However, in streaming scenarios, data distribution generally change through the time, that is, the data is non-stationary. The main challenge in such online environment is the model adaptation for the co...
Article
Concept drift refers to an alteration in the relations between input and output data in the distribution over time. Thus, a gradual concept drift alludes to a smooth and gradual change in these relations. It generates a model obsolescence and quality decrease in predictions. Besides, there is a challenging task: the extreme verification latency to...
Preprint
Full-text available
High-accurate machine learning (ML) image classifiers cannot guarantee that they will not fail at operation. Thus, their deployment in safety-critical applications such as autonomous vehicles is still an open issue. The use of fault tolerance mechanisms such as safety monitors is a promising direction to keep the system in a safe state despite erro...
Thesis
Full-text available
Em 2018, aproximadamente 82.000 pessoas desapareceram somente no Brasil. Entretanto, as bases de dados públicas existentes costumam ser desatualizadas. Além disso, ainda não existe uma base de dados unificada de todo país, provocando assim a existência de múltiplos registros com informações complementares ou divergentes sobre a mesma pessoa em base...
Conference Paper
Full-text available
Redes neurais artificiais (RNAs) são amplamente utilizadas, como por exemplo, em jogos digitais através de agentes inteligentes que replicam o comportamento humano. Contudo, o custo computacional de treinamento delas costuma ser alto, exigindo a escolha entre maior processamento ou menor qualidade. Desta forma, este trabalho propõe o uso de redes d...
Conference Paper
Machine learning (ML) provides no guarantee of safe operation in safety-critical systems such as autonomous vehicles. ML decisions are based on data that tends to represent a partial and imprecise knowledge of the environment. Such probabilistic models can output wrong decisions even with 99% of confidence, potentially leading to catastrophic conse...
Research
Full-text available
In data stream scenarios, the concept drift is a common problem. It refers to a modification in the relations between input and output data in the distribution over time, decreasing the accuracy of machine learning (ML) models. Besides, another frequent condition is when the instances’ labels are only available in the initial setting. This scenario...
Conference Paper
Full-text available
A distribution utility has to deal with several customer calls regarding grid maintenance or energy issues. Generally, when the proper channels receive a call from the customers, the reported issue pass through a screening phase and, in the end, a maintenance team is sent to the location to solve the problem. However, not all problems are responsib...
Conference Paper
Full-text available
The resource allocation problem is the process of allocating limited resources for a vast amount of tasks. Within this problem there are several important variants such as the stochastic time-variant resource allocation problem. This problem is relevant within environments where the distribution of resources varies with time, bringing difficulties...
Article
Full-text available
The resource allocation problem is the process of allocating limited resources for a vast amount of tasks. Within this problem there are several important variants such as the stochastic time-variant resource allocation problem. This problem is relevant within environments where the distribution of resources varies with time, bringing difficulties...
Conference Paper
Full-text available
Recommender systems generally are made to predict user preferences' for items. However, in high dimensional datasets this task demands high computational costs. Taking into account that data distribution changes through time, it is important that online recommender systems have a fast retraining process in order to keep the model updated, deliverin...
Conference Paper
Full-text available
Non-stationary environments suffers from a problem denominated concept-drift, that is, data distribution changes through the time. Solve such problems is extremely important in artificial intelligence field. Generally, algorithms that deal with concept-drift are applied in stream scenarios such as, robotics, image classification and sensors. Howeve...
Conference Paper
Full-text available
This article analyzes the performance of parallel programming in similarity metrics in the analysis of social networks, in order to identify the influence of individuals on the network. Thus, we propose a new metric, denominated Vertex Rank, whose objective is classify the influence of each vertex from its degree centrality and from its closest nei...
Conference Paper
Full-text available
Geospatial services have been used in several application areas. In this work, we create a mobile crowdsourcing tool aiming to assist people to publish , finding and recommend nearby services from small business. Our model is based on the concepts of Computer Supported Cooperative Work. The use of crowdsourcing for voluntary production of informati...
Conference Paper
Full-text available
In Brazil and around the world, Civil Society Organizations (CSOs) provide valuable public services for society. Through CSOs, people have organized and defended their rights, communities and interests, and can fully exercise their collective potential, often acting in partnership with governments to carry out public policies and/or develop their o...
Conference Paper
Full-text available
Increasingly, graphics cards (GPUs - Graphics Processing Units) have been used to increase the performance of computational algorithms, and, in the last years, several libraries have appeared in order to create a abstraction layer to developers in using GPUs. However, few studies evaluate the efficiency of these libraries in using GPUs. This articl...
Conference Paper
Full-text available
The more cities grow, the more transportation systems face increasing pressure. Although Brazilian citizens have mostly used public transportation as a solution, Brazilian public transportation systems present serious quality problems. There are several solutions to the problems of public transportation, however, few initiatives are focused on supe...
Conference Paper
Full-text available
Density estimation is a important statistical method that helps to determine, in a concise way, the probability of a phenomenom about certain types of data. Some kinds of data doesn’t have a known distribution but may reveal too much about certain domains, for instance, the geographic data. This work seeks to create an intuitive interface for geogr...
Data
Brazilian Missing people database containing 1794 registers from 7 different Brazilian datasources
Conference Paper
Full-text available
Resumo Estimar densidades sobre uma população é um trabalho importante e geralmente usamos uma função densidade para isso. A função de probabilidade é um conceito fundamental em estatística e existem diversas técnicas que podem ser empregadas para estimar uma população e a técnica escolhida para este trabalho foi o KDE. Porém, o KDE demanda um gran...

Questions

Question (1)
Question
i found some applications and some articles about it, but i haven't found any downloadable version nor API.
Has USA or any country, a database available with this information?
Kind Regards
Raul

Network

Cited By

Projects

Projects (5)
Project
The main objective of the Safer Autonomous Systems (SAS) project is to identify ways that we can establish people’s trust in autonomous systems by making these systems demonstrably safer. This project has received funding from the European Union's EU Framework Programme for Research and Innovation Horizon 2020 under Grant Agreement No 812788.