Quem eu sou
Eu possuo um PhD em Ciência da Computação pela Universidade Federal de Minas Gerais (UFMG), que eu completei em 2023 sob a orientação do Dr. Adriano Veloso. Durante o meu PhD, fui convidado a realizar uma visita técnica na Universidade de Stanford em 2022. Também possuo bacharelado e mestrado em Ciência da Computação pela UFMG. Antes de concluir meu PhD, trabalhei no LUAR-UFMG de março a dezembro de 2013, onde construí fluxos de trabalho e auxiliei no projeto Flux. Depois, trabalhei no LLP-UFMG de março a dezembro de 2014 no projeto de extração de benchmark orientado a loops. De fevereiro de 2015 a dezembro de 2018, trabalhei no laboratório J-UFMG. Em 2020, ministrei disciplinas de IA na graduação da PUC-Minas. Atualmente, sou membro desde 2018 do Laboratório LIA-UFMG, que se concentra em IA centrada no ser humano. Além disso, trabalho como líder técnico na Kunumi desde 2020.
Contato
Universidade Federal de Minas GeraisInstituto de Ciências Exatas
Av. António Carlos, 6627
Belo Horizonte - MG, Brasil
E: gzuin@dcc.ufmg.br
Publicações
Gianlucca Zuin, Rob Buechler, Tao Sun, Chad Zanocco, Daniella Castro, Adriano Veloso & Ram Rajagopal Revealing the Impact of Extreme Events on Electricity Consumption in Brazil: A Data-Driven Counterfactual Approach. IEEE Power & Energy Society General Meeting, 2022.
Qualis: A1
[Slides]
Gianlucca Zuin, Daniella Araujo, Vinicius Ribeiro, Maria Gabriella Seiler, Wesley Heleno Prieto, Maria Carolina Pintão, Carolina dos Santos Lazari, Celso Francisco Hernandes Granato & Adriano Veloso Prediction of SARS-CoV-2-positivity from million-scale complete blood counts using machine learning. Nature Communications Medicine. Nature, 2022.
Qualis: A1
Gianlucca Zuin, Felipe Marcelino, Lucas Borges, João Couto, Victor Jorge, Mychell Laurindo, Glaucio Barcelos, Marcio Cunha, Valdeci Alvarenga, Henrique Rodrigues, Paulo Balsamo & Adriano Veloso Predicting Heating Sliver in Duplex Stainless Steels Manufacturing through Rashomon Sets. IEEE International Joint Conference on Neural Networks (IJCNN). IEEE, 2021.
Qualis: A2
G. Zuin, L. Chaimowicz, & A. Veloso Deep learning techniques for explainable resource scales in collectible card games. IEEE Transactions on Games. IEEE, 2021.
Qualis: A3
G. L. Zuin, Adriano Veloso, João Cândido Portinari & Nivio Ziviani. Automatic Tag Recommendation for Painting Artworks Using Diachronic Descriptions. IEEE International Joint Conference on Neural Networks (IJCNN). IEEE, 2020.
Qualis: A1
[Slides] [Video-presentation]
G. L. Zuin & A. Veloso. Learning a Resource Scale for Collecfatible Card Games. IEEE Conference on Games (CoG). IEEE, 2019.
Qualis: A2
[Slides]
E. S. L. Zuin & G. L. Zuin. Computadores: aspectos históricos de uma invenção do século XX. 1ed. Belo Horizonte: PUC Minas, 2019, v. 1, p. 45-60.
Book chapter.
G. L. Zuin, L. Chaimowicz & A. Veloso. Learning Transferable Features For Open-Domain Question Answering. IEEE International Joint Conference on Neural Networks (IJCNN). IEEE, 2018.
Qualis: A1
[Slides]
A. R. Tavares, G. L. Zuin, H. Azpúrua & L. Chaimowicz. Combining genetic algorithm and swarm intelligence for task allocation in a real time strategy game. Journal on Interactive Systems (JIS). SBC, 2017.
Qualis: B3
G. L. Zuin, L. F. G. Magalhães & T. C. Loures. MAL-FITT: MyAnimeList Forum Interpreter Through Text. Encontro Nacional de Inteligência Artificial e Computacional (ENIAC). SBC, 2016.
Qualis: B4
[Slides]
G. L. Zuin, Y. Macedo, L. Chaimowicz & G. L. Pappa. Discovering Combos in Fighting Games with Evolutionary Algorithms. ACM Genetic and Evolutionary Computation Conference (GECCO). ACM, 2016.
Qualis: A1
[Slides] [Video]
G. L. Zuin & Y. P. A. Macedo. Attempting to discover infinite combos in fighting games. Brazilian Symposium on Computer Games and Digital Entertainment (SBGames), 2015.
Qualis: B2
[Slides] [Video-presentation]
A. C. Faria-Campos, L. B. Balottin, G. Zuin, V. Garcia, P. H. Batista, J. M. Granjeiro & S. V. Campos. FluxCTTX: A LIMS-based tool for management and analysis of cytotoxicity assays data. BMC bioinformatics, 16(Suppl 19), S8, 2015.
Qualis: A1
Projetos
Terranswarm
Implementation of the techniques described in the paper Evolving swarm intelligence for task allocation in a real time strategy game.
Basically, a Task Allocation algorithm based on swarm intelligence was employed to control a StarCraft bot. Then, a Genetic Algorithm was used to tune its parameters.
.
HMM-Combo
Implementation of the techniques described in the paper Attempting to find infinite combos in fighting games using hidden markov models.
Basically uses Hidden Markov Models and supervised learning to train a machine to find large combos inside fighting games. The consultations are made using the viterbi algorithm.
.
Loop Extraction
Utilizing the LLVM compiler and the pass framework, we seek to slice real programs in segments of code containing only a loop and the variables which affect its execution. The proposed extractor operates only over a program's ByteCodes, keeping the original program's integrity intact. In the next step, we translate low-level ByteCodes do high-level C programs.
.
.
Flux-INMETRO
This project aims to develop a system for data management that supports adherence of laboratories to the Good Laboratory Practices (GLP) and has among its features the hability to alllow traceability of test substances, with emphasis on the chain of custody and control of permissions of users; flexibility to suit different types of laboratory data, tools for identification of experimental features that best suit GLP; tools to provide a way for production of automatic or semi-automatic management reports that allow to monitor the implementation of this system of quality management.