Machine Learning

Dear students,

we have carefully considered all of your remarks and have adjusted some points where appropriate. The grades in CMS are now updated.

Your final grade will soon also be uploaded to LSF.

 

We hope you enjoyed the course and wish you the best for your future endeavours.

Best,

ML Team

Dear ML students,

We have finished grading the re-exam and have now published the points and grades.

 

Inspection:

As a reminder, the inspection for the re-exam will take place on Wednesday, October 18th, starting at 9am. Please be there no later than 10:30am and remember to bring your student ID card. The location is lecture hall 003 in E1.3.

 

Best regards,

ML Team

Dear ML students,

this is to remind you that the ML re-exam will take place tomorrow, 5th of October, at 12:30pm in E2.5, Lecture Hall 001.

The exam instructions can be found here: https://cms.sic.saarland/coreml/dl/72/Exam_Instructions.pdf

Please be there early and bring your student ID card!
 

We wish you good luck!

Best regards,
ML Team

Dear ML students,

your course grades have now been uploaded to LSF.

It is now also possible to register for the re-exam. The re-exam is on October 5, 12.30pm. Please register via LSF/HISPOS by September 28th. In case that registration via LSF/HISPOS is not possible for you, please send an email (with your full name, study program and matriculation number) to nweis@cs.uni-saarland.de before September 28th, so that we can register you manually.

No additional CMS registration is required for the re-exam.

Best regards,
ML Team

Dear students,

we have carefully considered all of your remarks and have adjusted some points. The grades in CMS are now updated.

Your final grade will soon also be uploaded to LSF. It will then also be possible to register for the re-exam. We will keep you informed.

 

The re-exam will take place on October 5, 12.30pm, as announced at the beginning of the semester. More details will be published soon.

 

Best,

ML Team

Dear ML students,

We have now uploaded your grades for both the exam and the project reports, and the resulting overall course grade (which includes a possible bonus from the project challenge).

On your personal status page, under "Tests and Exams", you will see, in this order:
- The grade of the exams overall, i.e. the best of main and re-exam grade
- The points and grade of the main exam
- The project report grade
- Project challenge bonus: If it says "+0.3", you got the bonus, if it says "unknown", you do not.
- Also, you can find the final project leaderboard in the notion link.
- Next, under "Gradings", you will see your final resulting grade for the course, as of now.

 

The exam inspection will take place on September 18th at 9am. More details will be shared soon.

Best regards,
ML Team

Dear ML students,

The results of the ML exam have now been shared with you via CMS. There you can see the total number of points that you got (out of 40) and if you passed the exam (passing threshold is 18 points out of 40). The final grade is still pending and will be share in the following week (also including the results from the Twitter project, when applicable).

The exam inspection will take place on September 18th at 9am. More details will be shared in the following weeks.

 

Best regards,

Prof. Dr. Isabel Valera (on behalf of the whole ML team)

 

 

Dear all,

We have now opened the following two submissions, both with deadline 17th of August:

Twitter Challenge REPORT: Please submit here the the pdf with report of the Twitter project. Please make sure to carefully read and follow the instructions, and use the provided template, also available via Overleaf.

Twitter Challenge CHALLENGE: For those participating in the challenge, remember to submit a .zip file with your predictions (both for the classification and regression tasks) for the 'tweets_test_2' dataset and make sure that you follow the correct format for your submission. Make sure to read the instructions also for this. In addition, we will update the leaderboard based on your predictions for the 'tweets_test_1' dataset three more times (Friday 11, Monday 14 and Wednesday 16), to make sure that you get enough feedback on both your submission format and estimated ranking for the challenge.

Remember that the deadline above is final and no additional submissions/corrections will be accepted after the deadline. May you have questions, please use the Forum.

The ML team hopes you enjoy the last push for the project and is looking forward to see your results and reports!

 

 

 

 

 

 

 

 

 

 

 

Dear all,

We have released test_2 dataset yesterday. It will be used for your final grade.

Please note that we have not opened a submission for this case. We will update you about this next week.

You can still submit your results on test_1 dataset for the leaderboard updates. We plan to update the leaderboards on Monday, Wednesday, and Friday from now on.

Best,

ML Team

Dear ML students,

Now that the registration for the exam is finally closed, we have uploaded the list of students registered for the exam to CMS. If you have registered via LSF/HISPOS or via email for the exam, please make sure that you appear as registered for the "ML Exam registration" in CMS. Otherwise, please let us know asap via email to nweis@cs.uni-saarland.de.

In addition, (as soon as possible and definitely before the exam) read carefully the instructions/rules for the exam shared https://cms.sic.saarland/coreml/dl/72/Exam_Instructions.pdf

Finally, for those participating in the exam, remember that the exam will take place next Thursday, 3rd of August at 12.30 hours at the GHH.

May you have further questions, please reach us out via the forum.

 

Best regards and see you soon,

Isabel Valera (on behalf of the ML team)

 

 

 

 

 

 

Dear all,

 

My most sincere apologies for skipping class yesterday but I was feeling unwell with a virus. I hope to be fully back on Thursday.

Remember that on Thursday (our last class), I will be answering your questions about the materials and exam. So please come with your list of questions ready. Also, tutors will be available to answer questions regarding exercises during this week tutorials.

 

Best regards,

Prof. Isabel Valera

 

 

 

Dear ML students,

The ML exam is approaching, as it will take place on August 3rd at 12.30 hours at the GHH (and if necessary at HS002 in E1 3).

In order to be able to take part in the exam, please register via LSF/HISPOS by July 27th. In case that registration via LSF/HISPOS is not possible for you, please send an email (with your full name, study program and matriculation number) to nweis@cs.uni-saarland.de  before July 27th, so that we can register you manually.

No additional CMS registration is required for the re-exam.

Best regards,

Isabel Valera

 

 

 

Dear all,

The surveys for the evaluation of the course, both lectures and tutorials, are now open for you to give your opinion on how the course is going. Please provide your feedback there by July 9th, as this will help us to further improve the course for following years (as results will only be available once the course is over). Use the following links to evaluate both the lectures and tutorials, respectively:

 

Machine Learning (lectures): https://qualis.uni-saarland.de/eva/?l=143170&p=1x5llc

Machine Learning (tutorials): https://qualis.uni-saarland.de/eva/?l=1431701&p=ec08gg

 

For more short term feedback, questions, etc. please do not hesitate to contact ML team via the forum or after the lecture/ tutorials. We are here for you!

Best regards, 

Isabel Valera (on behalf of the whole ML team)     

 

Dear all,

We have now updated the Twitter leaderboard to include also a list of the teams that had the wrong format in their submission. Please check there for available feedback on the wrong formatting.  If your Team ID is not in either the Ranking tables or the Feedback table, you may have a problem with your “.zip” file or some other error. In such a case, for further information please send an email to koyuncu(at)cs(dot)uni(dash)saarland(dot)de for a further check.

In addition, we have now reopen the submission for all teams*, such that you can submit at any time from now on. Please, notice however that we will only update the leaderboard (and thus give you feedback on your results and formatting) every Monday until August 14 (included).

Remember that final submission deadline for both predictions on test set 2 (which will be released on August 7th), as well as for the report will be August 17.

Best regards,

ML team

* Note: we decided to be lenient and allow teams that did not manage to submit on June 9, to still join the project.

 

Dear all,

After an initial check over the submissions, we would like to remind you couple of things:

• Make sure that you use the format {team_id}__{split}__{task}_pred.npy We realized some teams are using invite codes instead of Team ID. Moreover, your {team_id} should match in both of your submissions.
• You need to send your classification predictions without encoding, i.e. your predictions should consists of ['neutral', 'positive', 'negative']
• Please make sure that both of your submissions has the same shape with your input, i.e. given an test input shaped (1000,), your submissions should have the same shape.

If you have questions please use the forum to reach us.

 

Best,

ML Team

Dear all,

We have now updated the tentative schedule of the ML course. Some reminders:

- This week there are no tutorial sessions, but there is class tomorrow (Convex Optimization II)

- Next week there is no lecture but there are tutorials

- We have split  the exercise sheet of block 3 into two sheets and planned two solution tutorials for them.

- Remember that 09.06.23 is the checkpoint for submitting the checkpoint of the Twitter project necessary to start the leader board.

 

See you tomorrow in class,

Isabel Valera

 

PS. The material on convex optimization has been also updated/uploaded. Please download the new version of L10.

 

 

 

Dear ML students (especially those attending remotely),

Unfortunately, the Wifi of the building dropped and it was not possible to connect anymore. Please have a look to the materials of L09 and let me know if you have any questions. As a side note, the statistical test part is assumed to be advanced material and won't be part of the exam. The estimation of the performance and model selection based on that, it is indeed important. So I will do a recap on Thursday and take all your questions.

Sorry for the hectic lecture today and let's hope that Thursday goes better.

Best regards,

Isabel Valera

 

 

 

Dear all,

There has been a problem with an exam from another course that is on the same room as we had our lecture today. Since they do not fit in any other room, let;s move to our lecture to room 001 at E1.3.

 

Please come over, I will wait a bit longer for you here.

Thanks for your flexibility,

Prof. Isabel Valera

Dear all,

Now that you should have a general picture of the course content, as well as your semester load, I have open the course deregistration, so that students that decide not to participate actively in the course this semester can deregister for the course. This action, has two purposes: i) answer to those students that asked me to deregister them (now you can do by yourself); and ii) for the ML team to have an estimate of the number of active students.

In case you have decided not to take part of the course this semester, please deregister yourself from cms by June 9.

Thanks,

Prof. Isabel Valera

 

 

 

Dear all,

we decided to upload solutions for the first exercise sheet to CMS. They were handwritten by one of our tutors. If you have any questions about the solutions, please ask in a help tutorial or use the forum.

 

Whether we will upload (partial) solutions for the following exercise sheets is yet to be determined.

Have a good weekend!
ML Team

Hi all,

Tutorial of today (Thursday 11) at 2pm is happening as usual in room HS 003, building E 1.3.

As there is a public holiday next Thursday (May 18th), there won't be tutorial on that day. However, if desired, students assigned to the Thursday tutorial slot will be welcome to join the tutorial session on May 19th at 10am in room HS 002, building E 1.3.

Best regards,

ML team