Aprendizagem Automática

As aulas teóricas são muito interessantes. Penso que os slides podem possui mais alguma informação adicional de forma a perceber melhor matérias mais complexas. Teste acessível, no entanto, os projetos, que são muito interessantes, são pesados.

Estudar exames do ano anterior

Not much time for developing the lab assignments with very few support on how they should be done. Exam is not so difficult however.

I have to be frank and admit that I stopped attending theoretical classes early, they were not very exciting. The materials where ok in terms of adressing the concepts but I did not find them usefull for the lab assessments. The projects are progressively more difficult, although I believe that as long as you have a good lab teacher you will do just fine. I will have to admit that i did not find it particularly interesting. It's somewhat demotivating to take an exam where practically all the points are based on multiple choice, and reasoning isn't taken into account, but at least it doesn't...

Rather boring introduction to machine learning. Lots of powerpoints of algorithms with some theory, nothing amazing. Easy practical projects but with a strange evaluation method. The best results get the best grades so you are rewarded by exploiting the problem at hand for example, by fitting noise or guessing the number of outliers. Overall dull course, easy to complete but not much more.

Very interesting subject and a very good professor lecturing it. I alredy had interested for the topics of machine learning but this course dives into the basics of supervised learning in a very structured way.

It's a good course.

But you will only learn the basics. If you already have some knowledge, it can be a pretty easy course and at the same time boring.

In my opinion, nowadays it's a fundamental course for everyone studying engineering.

The professors are very accessible and nice and the workload isn't difficult. You can study at home and still get great grades.

Laboratory classes

There are 3 projects and you can complete each one in just a day. However, the project grades are usually standard (15-16).

Good course that covers the fundamentals in machine learning. The projects are to be done in teams of 2 members, and the objective is to implement some ML algorithm and benchmark it against the dataset specified. The project can be a bit difficult but the grades are generally good.

The projects are a bit adhoc and take some time to do. You can try a lot of models and choose the best and still ending up with a 15/20. The profs are good, the content is important.

I think this is a good introduction course to machine learning. However, the classes where not amazing. The theoretical classes were the definition of mid and I think you will be okay by skipping them and reading the material by yourself. The problem classes consist on a rush of solving exercises on the board. If you have someone who can provide you the solutions to the problems, I feel like there is no point on going to those classes. The lab classes are mandatory and are where you do the project. I learned a lot while doing the project, however I think the evaluation method is not the best....

Good introductory course for ML. Heavy workload since we had to deliver a homework every week. The slides were good. Tip: go to exercises classes. Teachers are good for all classes: lectures, exercises and labs.

The course is very pratical, hands-on in real-life problems. Professor Catarina Barata will help you with the project and if you attend the lab and problem classes you shouldn't have problems with the project nor the exam.

Evaluation with almost weekly projects and exam was too much. Also multiple choice exam is not the best way to evaluate the taught contents.

Meh.

It gives a decent foundation for Machine Learning and if you're a beginner I'd recommend it. But honestly, I'd argue Deep Learning is a much better course that basically talks about all of the subject matter in Machine Learning (except maybe Random Forest which can be easily learned online).

The project component consists of 4 projects for a 7 week course, which is really overkill (doesn't help that the labs are essentially just for questions regarding the project, and you pretty much have to figure how to use stuff on your own). The regression problems (first two) were alright. But the...

Essential course for anyone who wishes to know something of Machine Learning. Overall well structured and good evaluation

Mandatory foundational course for the AI world, choose it as fast as possible. The project is interesting but needs constant work. The exam is not too difficult.