May 2018 - mllab@ins.uni-bonn.de

[Mllab] [INS Website][Feedback mllab] Task 3.4 garb_train_…
by INS Website 31 May '18

31 May '18

Hello, new feedback arrived for mllab: Task 3.4 garb_train_25x50 used twice. Probably one has to be replaced by test. Metadata: vcs: d1a4117c page: 32 obj: Script misc: referer: Have a lovely day! Your Website

1 0

[Mllab] Practical Lab
by Leon Tim Engelbert Sievers 28 May '18

28 May '18

Sehr geehrte Praktikumsleitung, ich nehme zwar offiziell nicht an Ihrem Practical Lab teil, interessiere mich aber dennoch dafür, Anwendungen im Bereich Machine Learning kennenzulernen. Die ersten beiden von Ihnen veröffentlichten Übungsblätter habe ich bereits bearbeitet und fand sie sehr motivierend und nützlich in Bezug auf dieses Bestreben. Nun würde ich gerne auch das dritte Übungsblatt bearbeiten, habe aber das Passwort für die erforderlichen Daten nicht mitbekommen. Wäre es möglich, dass Sie mir dieses per Mail zukommen lassen? Mit freundlichen Grüßen, Leon Sievers

3 2

[Mllab] [INS Website][Feedback mllab] Sweet…
by INS Website 20 May '18

20 May '18

Hello, new feedback arrived for mllab: Sweet Metadata: vcs: d1a4117c page: 33 obj: Script misc: referer: https://beta.ins.uni-bonn.de/media/public/courses/SS18/practical-lab-numeri… Have a lovely day! Your Website

1 0

[Mllab] Regarding sheet 3, task 3.3c
by Kin Tung Jonathan Chan 18 May '18

18 May '18

Dear MLLab, I am not quite sure what task 3.3c is asking. This is my understanding of how we build the classifier: for the training phase, we apply PCA first to project the given data, and in our case this means the 2D and 1D PCA to prepare the data to be classified as red/not red and green/not green respectively. Then we train to SVMs for red/not red and green/not green. For the actual classifier itself, the data is projected with with the red PCA matrix and the green PCA matrix. These processed data are given as input to the red and green SVMs respectively (in a sequential manner). Then the result is constructed in the obvious way based on the SVM results. However, I am particularly confused by the comment “print accuracy on 1, 2, 4 dimensions”, for based on my understanding, we are supposed to use at least two different dimensions in our classifier. Jonathan

2 1

[Mllab] [INS Website][Feedback mllab] nice pictures…
by INS Website 17 May '18

17 May '18

Hello, new feedback arrived for mllab: nice pictures Metadata: vcs: d1a4117c page: 33 obj: Script misc: referer: Have a lovely day! Your Website

1 0

[Mllab] [INS Website][Feedback mllab] an orthogonal transf…
by INS Website 16 May '18

16 May '18

Hello, new feedback arrived for mllab: an orthogonal transformation can also be a reflection(I am referring to the last line) Metadata: vcs: d1a4117c page: 29 obj: Script misc: referer: Have a lovely day! Your Website

1 0

[Mllab] [INS Website][Feedback mllab] I think it would be …
by INS Website 16 May '18

16 May '18

Hello, new feedback arrived for mllab: I think it would be more appropriate to say that the columns of V_q form an orthonormal system, since they dont span R^d unless q=r Metadata: vcs: d1a4117c page: 27 obj: Script misc: referer: Have a lovely day! Your Website

1 0

[Mllab] Solution for Sheet 2 [MachineLearning] Juleana, Xinda
by Xinda Wu 16 May '18

16 May '18

Hi Dr Bastian Bohn, In the attachment is the solution from our group. Best, Xinda

1 0

[Mllab] sheet2
by Janos Meny 16 May '18

16 May '18

To whom it may concern, i have attached franziskas and my solutions for the second exercise sheet. Best regards Janos

1 0

[Mllab] Sheet 2
by Adolfo Camacho Melo 15 May '18

15 May '18

Anbei die Lösungen des zweiten Blatts. Grüße Adolfo

1 0