Research Themes
Research
From deep learning to biometric recognition — our research groups cover the main frontiers of intelligent computing.
01Research Areas
Fundamental Methods, Real Problems
Each area brings researchers together around a set of methods and concrete questions from healthcare, finance, industry and the public sector.
-
Artificial Intelligence
AI models applied to high-complexity problems in healthcare, finance, language and environmental perception.
Topics
- Artificial Intelligence
- Computational Intelligence
- Applied Machine Learning
-
Computer Vision
Visual pattern recognition, image classification, deepfake detection and biometric identification.
Topics
- Computer Vision
- Biometrics
- Periocular Biometrics
- Pattern Recognition
-
Machine Learning
Deep neural networks, transfer learning, self-supervised learning and efficient architectures (NAS).
Topics
- Machine Learning
- Deep Learning
- Metric Learning
- Applied Machine Learning
-
Natural Language Processing
Language models for Portuguese, toxic text classification, speech recognition and prompt optimization.
Topics
- Natural Language Processing
- Text Classification
-
Data Science
Credit risk analysis, fraud detection, churn prediction and epidemiological time series.
Topics
- Data Science
- Statistical Methods
- Time Series
- Databases
- Network Science
-
Spatial & Epidemiological Analysis
Spatial cluster detection, COVID-19 modeling with GNN and urban mobility, epidemiological surveillance.
Topics
- Spatial Analysis
- Cluster Detection
Researchers
-
Biomedical Signals
ECG classification, arrhythmia detection, heartbeat segmentation and remote cardiac signals via video (rPPG).
Topics
- Biomedical Signals
- Biological Signal Analysis
-
Multi-Objective Optimization
Evolutionary algorithms, community detection in social networks and Pareto solution dispersion/concentration optimization.
Topics
- Multi-objective Optimization
- Evolutionary Algorithms
- Community Detection
- Social Network Analysis
02Technology Transfer
Consulting & R&D
Applied research projects with industrial and government partners.
-
EFI Bank
OngoingAI for the Financial Sector
3 active R&D projects in partnership with EFI Bank. We develop AI models for deepfake detection, fraud detection, credit analysis and customer management. Led by Prof. Gladston Moreira and Prof. Eduardo Luz.
-
CGU
NLP Models for the Plataforma Brasil
Development of Natural Language Processing models for the Comptroller General of the Union (CGU), applied to the Plataforma Brasil for automated analysis of government documents and processes.
-
Embrapii / Escola de Minas
Industrial AI
In partnership with Embrapii and UFOP's Escola de Minas, we develop AI models for industrial applications, bridging fundamental research with national industry demands.
-
BDMG
Data Science for Credit
Consulting in data science and credit risk assessment for the Minas Gerais Development Bank (BDMG), developing predictive models for portfolio analysis.
-
BDI
In developmentCredit Risk
Project in development focused on data science for credit risk with BDI, expanding our activity in the financial sector with advanced machine learning techniques.
03Open Science
A Lab That Understands Innovation
Our projects stem from real-world questions and produce results published in the top international venues — IEEE, Springer, SBC — along with open-source code available to the community via GitHub.
14
public repositories on GitHub
312
publications on Google Scholar
6,164
citations
- 2015: 16
- 2016: 15
- 2017: 22
- 2018: 19
- 2019: 9
- 2020: 17
- 2021: 25
- 2022: 32
- 2023: 25
- 2024: 36
- 2025: 53
- 2026: 43 (year to date)
Collaborate with CSI Lab
Companies, public agencies and research groups can work with us on R&D projects, consulting and joint research. Tell us about your challenge.