Este Workshop visa auxiliar profissionais da saúde a entenderem como o ChatGPT e outras ferramentas de Chatbot podem ser utilizadas como aliadas no seu dia-a-dia de trabalho, estudo e atividades acadêmicas. O workshop abordará informações essenciais como usar comandos no ChatGPT e no Bing AI para auxiliar em atividades de aprendizado, pesquisa e assistência clínica. Além disso, os participantes vão entender as principais potencialidades e limitações do uso destas ferramentas aplicadas ao setor da saúde.
Fui entrevistado pela Folha de São Paulo sobre aprendizado por imitação (tema de pesquisa que abordamos em conjunto com veículos autônomos), e sobre a técnica usada no bot de minecraft da OpenAI. Apesar da matéria ter ficado imprecisa em relação a minha entrevista, o link se encontra abaixo:
Master student Irving Petrazzini, supervised by me, proposed PPO with the Beta distribution in 2021, surpassing the state-of-the-art on the CarRacing environment from OpenAI Gym. Check our paper for more details:
In terms of research, I will continue investigating Machine Learning and Neural networks approaches mainly for intelligent robotics, and data-driven modeling in oil and gas
From June 20 to June 23 (2018), I am pleased to attend the ICAPS summer school in Noordwijk, the International Summer School on Planning and Scheduling.
Invited by prof. Thorsteinn Rognvaldsson, from December 4th to December 8th (2017), I have given a project-based short course on Echo State Networks (Reservoir Computing) for PhD students/postdocs at Halmstad University, Sweden.
After some period in the Deparment of Automation and Systems at Federal University of Santa Catarina, and the Department of Mechanical Engineering at the same university, I have joined as a research associate the SnT - Interdisciplinary Centre for Security, Reliability and Trust - at the University of Luxembourg. There, I will initially work in the project with Choice Technologies Holding, devising machine learning methods for fraud detection in energy distribution networks.
I have given a presentation on "Dynamical neural networks for data-driven soft-sensor modeling (in oil wells)" showing the lastest research on creating dynamical models (Echo State Networks) with real-world oil well data, during the V Oil and Gas Production Optimization Workshopat CENPES - Petrobrás, Rio de Janeiro, the largest research center in Latin America.