Anúncios

Change in part 1 of the project

11 outubro 2020, 17:43 Joao Manuel de Freitas Xavier


Dear all, 

We updated part 1 of the project. We changed a threshold in Task 1, part (c); before, we had || u(t) - u(t-1) ||_2 > 1e-6; now, we have || u(t) - u(t_1) ||_2 > 1e-4.

We changed the threshold to avoid numerical issues that some of you had already encountered.

Best wishes,
jx


Colab notebook demo

7 outubro 2020, 15:23 Cláudia Alexandra Magalhães Soares

Hi everyone,
Here is the colab notebook link for the python demo:


https://colab.research.google.com/drive/1Ywns__MfhsNa9YykrNjZp9Hk9B-doC2I?usp=sharing

Please authenticate with your Tecnico login.



Beginning of the project

1 outubro 2020, 18:08 Joao Manuel de Freitas Xavier

Dear all, 


The list of groups for the project is available: see section "Files for downloading."

Part 1 of the project is also available. Recall that the project has three parts. Parts 2 and 3 will be posted as scheduled in the slides of the course overview (see the last slide).

jx


Project support classes start tomorrow

29 setembro 2020, 14:21 Cláudia Alexandra Magalhães Soares

Dear all,

Tomorrow, Wed 30, we will have two project support classes, covering:

9:30 am --  the setup of CVX for MATLAB


I assume you will have a working MATLAB installation. 


Remember, if you decide to use MATLAB, you can install it via a campus license (check https://si.tecnico.ulisboa.pt/en/software/matlab/).


Please, test if your installation is working!


You should also download the free CVX package from http://cvxr.com/cvx/download/

I recommend watching the 15-minute tutorial online: https://youtu.be/N2b_B4TNfUM


2 pm -- the setup of cvxpy for Python and the Google Colaboratorium platform


I assume you will have a working Python 3.x distribution. 

I strongly recommend Anaconda (https://www.anaconda.com/products/individual), even if your system already has some python version.


Please, test if your installation is working!


After installing Anaconda, install cvxpy: https://www.cvxpy.org/install/index.html

Check 
https://youtu.be/9sDidkln7R0?t=2306

https://youtu.be/PFVOTBQJvSE?t=4940

about cvxpy and CVX, and (Convex.jl) presented in one of the editions of the super famous Machine Learning Summer School at Tübingen.

If you decided to use Python, the easiest way to use Python in this course is through Google Colaboratorium (https://colab.research.google.com).

For Google Colab, even if you don't have a google account, you can enable access through your Tecnico account. 

Check instructions at https://si.tecnico.ulisboa.pt/en/servicos/servicos-externos/google-suite-for-education/


The zoom link for all project support classes will be sent by email. So check your inboxes for the link, and see you tomorrow!



Group matchings for the project

27 setembro 2020, 11:44 Cláudia Alexandra Magalhães Soares

Dear students,


Some of you have asked if there was a communication channel to create group matchings for the project. I would suggest posting on the Fenix forum associated with the course. You may access it through your Fenix profile.