Welcome to the CSILab - Laboratório de Computação de Sistemas Inteligentes da Universidade Federal de Ouro Preto
If you wanna be a student or collaborator in the CSILab, here you will find information about how we work and tons of interesting projects to join in.
The fields we work in are:
- Artificial intelligence (Inteligência Artificial)
- Network science / Complex networks / Complex systems (Redes Complexas)
- Multiobjective optimization (Otimização Multiobjetivo)
- Machine Learning / Deep Learning (Aprendizagem de Máquina)
- Data Science (Ciência de Dados)
- Computational Statistics (Estatística Computacional)
- Computer Vision (Visão Computacional)
Professors:
- Eduardo Luz (eduluz@ufop.edu.br, Lattes CV, Website)
- Gladston Moreira (gladston@ufop.edu.br, Lattes CV, Website)
- Jadson Castro (jadson.castro@ufop.edu.br, Lattes CV)
- Pedro Silva (silvap@ufop.edu.br, Lattes CV)
- Rodrigo Silva (rodrigo.silva@ufop.edu.br, Lattes CV)
- Valéria Santos (valeriacs@ufop.edu.br, Lattes CV)
- Vander Freitas (vander.freitas@ufop.edu.br, Lattes CV, Website)
- Ivair Silva (ivair@ufop.edu.br, Lattes CV)
Find below some general guidelines for our students:
-
Planning: Effective planning is crucial. Create a comprehensive schedule with attainable goals and maintain it in a spreadsheet format. Maintain open communication with your supervisor through regular meetings, ideally on a weekly basis. Your supervisor may oversee multiple students, so it is imperative that you keep your progress updated. If needed, reach out to your supervisor via email or a personal visit to ensure the progress of your schedule is well-monitored.
-
Git: Before everything, create a git project named 'FirstNameLastName' with 2 folders: 'code' (for your code), and 'thesis' (for your latex files). We also encourage everyone to use overleaf for writing. If possible, sync overleaf with the git repository. Keep everything in your git repository up to date.
-
Experiment tracking: Organize your git repository. We strongly recommend using an experiment tracking tool such as MLFlow and WandB. Read this great MLFlow tutorial by Felipe Melo or if you prefer run this WandB example for a nice Dashboard for your experiments. Also, take time to read this guideline to help you organize your directories and experiments. Example of how to organize your git repository: link.
-
CSILab servers and GPUs: We have nice computers in our laboratory (128Gb RAM, >40 cores), with powerful GPUS (RTX 3090, etc), available for you to run your experiments (don't stress your laptop!!). Access to the lab machines is done remotely and managed by Prof. Pedro Silva (silvap@ufop.edu.br). Drop an email to Prof. Pedro (with a copy to your supervisor) for instructions on how to use our infra-structure.
-
Be an active member of the lab: Participate, use our physical space, meet other students and the professors. Live the lab! The exchange of knowledge is very important! Ask your supervisor to add you to the Lab's WhatsApp and Slack groups.
-
Working with Deep Learning: Training deep learning models is a difficult task. Finding good hyperparameters and architectures can be a tricky activity for beginners. If your model does not converge or is performing poorly, before going to your supervisor, take a look at the tutorials: Deep Learning Tuning Playbook, and A Recipe for Training Neural Networks. Mandatory reading: How to avoid machine learning pitfalls: a guide for academic researchers. See also Deep Learning Tuning Playbook,
-
Don't reinvent the wheel! Start from good code repositories. Get to know PaperWithCode.
-
How to write good papers: As a researcher or scientist, writing is a critical component of your work. To aid in this endeavor, we highly recommend following the guidelines set forth by Professor Jan Van Gemert. Our laboratory places great emphasis on adhering to these guidelines for all research papers, dissertations, and theses. It is recommended to thoroughly review these guidelines multiple times for optimal results. Also read this excellent guide on how to organize your article How to ML Paper - A brief Guide. Additionally, familiarizing oneself with well-written papers is essential to hone your writing skills. We suggest regularly reading high-quality research papers to gain insights and improve your writing style (well-written paper). For further reference, we also recommend the article "Heuristics for Scientific Writing (a Machine Learning Perspective)" as a useful resource in enhancing your writing abilities. Good tools to help with the literature review are Litmap, citationgecko and SCISPACE. Read about the Five ‘C’s while writing a literature review. For an example of a good paper, with solid methodology and literature review, see this article on time series forecast. Recommended reading Stanford Tips for Writing Technical Papers and Tips for writing NLP papers .
-
Use a template to start you paper: The Prof. Jan Van Gemert also has a great skeleton guidelines on how to write a paper. Here is an example of an article skeleton from Prof. Gemert.
-
Stuck? Meet your supervisor more often! We encourage you to engage in regular and open communication with your supervisor. Our goal is to assist you in every way possible, and regular meetings can help facilitate this support.
-
For PPGCC students: Get to know the program's regiment and maintain open communication with your student representative. Ultimately, it is your responsibility to manage and stay informed about your academic progress. Follow the PPGCC website and the social media.
-
For undergrad students: Pay attention to IC deadlines and check the PROPPI website frequently.
-
PPGCC/UFOP template for your TCC, Thesis and Dissertation.
Find below lists of projects you could get involved in according to your level (undergraduate, Master's degree, Ph.D.):
Undergraduate - Scientific Initiation [IC] and Honours thesis [TCC]
The CSILab – Intelligent Systems Computing Laboratory at the Federal University of Ouro Preto is a research group dedicated to advancing artificial intelligence, machine learning, computer vision, pattern recognition, and data science. Our mission is to foster innovation, develop cutting-edge technological solutions, and provide specialized AI consulting. By combining excellence in academic research with practical applications, we tackle real-world challenges and drive scientific and technological progress. Join us at CSILab, where ideas are transformed into impact.
O CSILab - Laboratório de Computação de Sistemas Inteligentes da Universidade Federal de Ouro Preto é um grupo de pesquisa dedicado ao avanço da inteligência artificial, aprendizagem de máquina, visão computacional, reconhecimento de padrões e ciências de dados. Nosso foco está em promover a inovação, desenvolver soluções tecnológicas de ponta e realizar consultorias especializadas em inteligência artificial. Combinamos pesquisa acadêmica de excelência com aplicação prática para enfrentar desafios reais, contribuindo para o progresso científico e tecnológico. Venha conhecer e colaborar com o CSILab, onde transformamos ideias em impacto.