Agentes Autónomos e Sistemas Multi-Agente

I disagree slightly on the evaluation methods in this course. The laboratory component is basically freestyle, we pick a theme/scenario involving agents and then have to implement it. Early on we are quite limited because we know nothing about agents nor agent orchestration (at least I did not) so I felt quite lost during the development, to be frank. Some themes might be "easier" that others and the laboratory teacher clearly had his favourites. The lab classes were 3 hours and were solely dedicated to either project support or the mandatory delivery presentations, but none felt useful enough...

As aulas teóricas são um pouco monótonas. O projeto teve três fases e em cada uma delas tínhamos que preparar uma apresentação para mostrarmos a evolução do projeto. Apesar destas três fases, em nenhum momento tivemos feedback sobre a nossa nota (apenas no fim do período) portanto não havia muito espaço para melhorar o projeto com dicas dos stores ao longo do tempo. Os materiais dos anos passados para estudo de exame eram acessíveis no entanto achei o exame em si com uma dificuldade muito acima desses mesmos materiais. Em 25/26 podemos levar uma folha A4 com notas escritas à mão para o exame.

Lectures felt somewhat repetitive and not very well organized, and some materials recycled from previous years had the background for a certain thing but then would not define them properly or how to use them (e.g coordination graphs).

Open-ended project, the labs teacher appeared for only 15 minutes of 3 hour long labs several times and provided little support.

This course was good. The project was really open for us to research and explore, which is good because the project ended up being really interesting and in a subject we like. We had to do 3 presentations: an initial presentation for the idea and plan of the project, a presentation showing some progress and a final presentation showcasing the project and results.

The big thing I didn't like about the course was that the presentations seemed kinda useless because we didn't get any grade or useful feedback from them. It's more for the professor to understand the topic. And since these...

Sugestão: Encontrei um bom grupo de trabalho e a cadeira faz-se tranquilamente, se tiverem um grupo desorganizado, é uma cadeira que custa muito a passar. Não é que a cadeira seja dificil, mas é suposto haver bastante pesquisa sobre as decisões que tomamos no projeto e defender as mesmas.
Os professores sã bastante simpáticos, mas a organização da cadeira é um bocadinho estranha. As aulas práticas são praticamente 2/3 horas para trabalhar em conjunto no projeto e pouco mais...
O exame é fácil e descobrindo o padrão das perguntas é fácil tirar uma boa nota.
Temas interessantes, mas a parte...

The AASMA curriculum focuses on Game Theory and some Reinforcement Learning, making it a good pairing with PADI (with some overlapping content).

The theoretical lectures, taught by Professors Rui Prada and Manuel Lopes, are okay. In those lectures, teachers go through the slides and solve some exercises, whose solutions are not provided on Fénix. None of the material is particularly difficult to understand, and the course provides a solid introduction to the field.

The project is open-ended, which introduces considerable variability. You can go from implementing simple grid...

The project is open-ended (you propose it yourself following certain rules set by the professor). This means you can choose whatever difficulty level you want. Since I wasn't very interested in the course, I chose a simple AI topic (basically adapted the lab exercises to an existing environment). Overall, adapting the environment took the most work because it had bugs, but honestly, it was one of the less demanding projects I've done. I got a 17/20 on the project.

For Professor Sardinha's lectures, I only attended about 50% due to scheduling issues. He doesn't add much beyond what's in the...

I took this course with Professor Rui Henriques (god bless his soul), but unfortunately he's not teaching it this year, so my feedback might be completely off.

The subject is super fun and interesting! It gives you a really cool perspective on how different types of autonomous agents work, interact, and make their decisions. The lectures are generally fun, with some probability and statictics, but nothing too complex.

The real gold of this course is the project: You can choose any topic you want, then you have to implement it and write a report about it. You can either make a simple single...