Aprendizagem Profunda

Definitely the course I have taken with the most amount of available materials!

The lectures were interesting and the assessment was adequate. The homework assignments took some time but were easy to solve with the given materials

Homework assignments can be easily completed using the material provided by the professors. The exam has some complex questions, but since it's open-book, it's possible to get a satisfactory grade even if you don't do well. The professors were committed and interested in answering questions. I'll leave a tip unrelated to the subject matter – besides Fénix, the Piazza platform was used as a means of communication between students and professors. I advise you to always pay attention because the information relevant to project discussions was only published on this platform, and several students...

Very complete. The practical exercises and projects are really hands on! Do really recommend.

Deep neural networks, transformers, CNNs, RNNs etc... It covers the principles of these and other models, the final exam is with consultation on notes and paper you want to take take just paper, it's is very interesting as you start to comprehend how LLMs work and classifications of images etc... The professor is straightforward and easy to understand the class.

This course is essential for anyone looking to build a strong Machine Learning and AI foundation and it is the natural followup of the Machine Learning course. While the lectures are high quality, they are quite dense. There are a lot of things to learn in a very short period of time and that's what makes the workload somewhat heavy. Besides the exam, the practical homework problems mimic real-world applications and are excellent for consolidating your understanding of the core topics.

Contrary to what most people write in the reviews, I find this course harder/more time consuming than the Machine Learning one. The first 3 weeks is just revision of the machine learning course (AAut) and then you delve into some specific deep learning architectures (RNNs, Transformers, CNNs). The course is well structured and the projects are interesting, but there is lots to learn! The exam was open book, and it is not very difficult.

Essential course for deep learning. It covers a lot and the projects are good to learn. Then there is an exam to with consultation. The teachers both from the lectures and the exercises class are very good.

The theoretical lectures are well structured and worth attending. The lecturer (Prof. Mário Figueiredo) presents the concepts in an interesting manner, and keeps the class engaging. The pratical lectures are crucial for staying up to date with the project and asking questions. The workload is significant, but manageable. The project requires a true understanding of the concepts. The exam, while it covers a lot of concepts, is fair.

The theoretical courses are important to be able to follow the slides that go along with the course. For the exams the slides and some complementary knowledge about some topics might be needed to answer the theoretical questions that they have

Great course to take after Machine Learning, it's a natural second step for the AI and robotics master's students. You will talk about neural networks but also about attention, transformers, and some state-of-the-art algorithms. The professors are nice and helpful.

Important course for the Data Science curriculum. Sometimes a little bit too mathy, but gives you a solid foundation in Deep Learning, from the beginning to the state of the art