Artificial Intelligence

Dear students,

we just published the results of the first exam. This is the grading scheme:

1.0     >= 90 points

1.3     85 ≤ points < 90

1.7     80 ≤ points < 85

2.0     75 ≤ points < 80

2.3     70 ≤ points < 75

2.7     65 ≤ points < 70

3.0     60 ≤ points < 65

3.3     55 ≤ points < 60

3.7     50 ≤ points < 55

4.0     45 ≤ points < 50

5.0     < 45 points

Best regards,
Daniel Höller

Dear students,

if you attend the first exam tomorrow, please read this news carefully. It includes many rules and organizational information that will be important to smoothly run the exam.

Pre-exam/general information:

• The exam will take 2 hours.
• The exam will have enough empty pages to write the solution on, no need to bring your own paper.
• You are not allowed to use a pocket calculator (nor any other electronic device).
• Remember to bring your student ID.
• Start time is 2pm, the rooms will be open from around 1:45pm. Please be on time!
• Make sure you know the room and seat that you are placed in. You can find this information on your CMS status page.
• Please do not form big crowds in front of the exam room entrances to ensure safety distance.
• We ask you to wear a mask when entering the exam room.

During the exam:

• We will check your student IDs before the exam will actually start and again ask you to have your mask on during this process, so it hopefully leads to less distraction during the exam.
• In case you have a question, or need to go to the washroom during the exam, please signal (raise your hand) and wait until we approach you. In both cases, we ask you to wear your mask.

After the exam:

• When the time is up, stop writing immediately, and keep seated. If you keep writing, the page of your exam you were still writing on will not be corrected.
• We will ask you to come to the front one-by-one (including all your belongings!), please wait for your turn. You will drop the exam in a box and leave the room.
   

All the best,
Your AI Team

Hi all,

some information about the lecture today (July 21):

This lecture is not exam relevant. What I aim to do is give you a rough impression of ongoing research in my group: Assuming we have learned a neural network that takes action decisions -- how can we gain trust in these decisions? We are currently pursuing approaches towards explanation, safety verification, failure testing, and visualization. I will briefly cover each of these at an abstract level, elucidate the main ideas, and point to open questions. We are always looking for students in these areas, so this could be your chance for a student assistant job, BSc thesis, MSc thesis (provided you are qualified which involves at least a very good grade in this course, and ideally a very good grade in the AI Planning course in the winter term 22/23).

The lecture will be in zoom only, given the currently quite high infection numbers.

best,

Jörg Hoffmann

Dear students,

Please check the CMS calendar for next week.

Lecture on Monday
There will be no lecture on Monday. The lecture on "Trust in Neural Action Policies" will be on Thursday.

Additional Tutorials
We decided to offer a couple of additional tutorials in the last lecture week in favor of the "exam preparation lecture" as a preparation for the exams. Each tutorial will cover a subset of the topics.

In these tutorials, we will present the solutions of additional exercises covering the respective topics. After that there is time for you to ask questions about the topics covered in the tutorial.

To be able to better organize the tutorials, we ask you to post your questions in the new category of the forum (Additional Tutorials). We have already added one thread per topic in the category. Please only use this thread to ask your questions.

These tutorials will be online-only. You will find the links to the tutorials in the CMS calendar in due time.

Best regards,
Daniel Höller

Dear students,

The CMS registration for the first exam is now open. As soon as you fulfill the requirements, i.e. you passed 8 out of 10 theoretical sheets and 3 out of 4 practical sheets, you can register yourself for the first exam on your personal status page. The CMS registration is open until July 19.

Note that you have to register for the exam in CMS, otherwise you cannot participate! This does also hold e.g. for the Erasmus students etc.

In addition to the CMS registration, you might have to register for the exam at your respective examination office (Prüfungssekretariat). For most of you, this should be the HISPOS registration. If you are unsure about this point, please contact your examination office.

The exam will take place on July 27, starting at 2pm, on campus, distributed across several lecture halls. We will inform you in advance regarding the details about how exactly the exam will be organized.

Kind regards,
Daniel Höller

Dear students,

we have spotted and fixed a mistake in practical sheet 3, exercise 2. In particular, constraint 4 now says:

4. Exactly one store has all 3 types of items snacks.

Additionally, we made a clarification regarding property (b):

(b) there is (at least) one snack sold by both Store 2 and 3

Please be aware of these changes. The exercise instruction has been changed accordingly in the current CMS revision.

Regards, Thorsten Klößner

Dear students,

I apologize for so abruptly ending the video stream of the lecure on Monday, but, unfortunatelly, my laptop suddenly died. In the last 10 minutes of the lecture we talked about some basic compilations of PDDL tasks. You can find short notes on this part on the last page of the whiteboard pdf in CMS. However, this part will not be part of the final exam, and I'll go over it again quickly at the beggining of the next lecture, which will be on Monday June 20 due to the public holiday this Thursday.

The remaining three lectures on planning will continue to be held in a hybrid form, but if you want to ask any questions you need to come in person to the lecture hall.

Best,

   Dan Fiser

ps. I downloaded the video, and uploaded it again. Seems to work. New link is in materials.

 

Dear all,

YouTube sent me this email today:

"

Our team has reviewed your content, and we think you may need to make changes to make sure it doesn't violate our spam, deceptive practices and scams policy. In the meantime, we've made the following content private:
 

Video: AI 2022 -- Lecture 11
 

We haven’t applied a strike to your channel, and if you make changes, or if you think we've made a mistake, you can request to make your content public again. However, there may be some instances where YouTube keeps your content private.
 
What "locked as private" means
 

Content locked as private won't appear on your channel or in search results, and it won't be visible to other users. Learn more about how this works.
"

So, if you're wondering why you cannot access the video, that's why.

What can I say (other than rolling my eyes). I'm trying to fix it.

best,

Jörg

Dear students,

starting next week, we will provide office hours where you can discuss practical sheet assignments. The office hours will be conducted online in MS Teams [1]. In order to attend, you need to register to the team called "[136388] Artificial Intelligence (Summer 22)" which you can do using the code nt2o4w9. If you are unsure how to use the code, following the instruction described here: https://support.microsoft.com/en-us/office/use-a-link-or-code-to-join-a-team-c957af50-df15-46e3-b5c4-067547b64548

All office hours were added to the CMS calendar already, so to attend office hours you can just follow the provided link. The meetings are also added to the MS Teams calendar, and you can see them listed in the "General" channel of the aforementioned team.

The next office hours will be on Tuesday May 17 at 16:15 with Eva Röper where you can ask questions about z3 solver.

Best,

   Dan

Hi all,

it may be that some of you did not find the lecture zoom link this morning as you expected it to be in the Timetable.

Sorry, I added the link to my lecture events now. In any case, the link is always the same:

https://cs-uni-saarland-de.zoom.us/j/98245956529?pwd=aU1oZ1ZUTGlxaEorUTNOR2FEU3Jhdz09

This is true for the entire course, including also those lectures we will do hybrid in GHH and zoom.

best,

Jörg

ps. On the side: I had forgotten to upload the CSP pre-handouts prior to the lecture. I have now uploaded all pre-handouts for the lectures up to June 9.

Hi all,

this is just to remind you (see previous news item below, sent on May 6) that the next edition of the AI course will be in the WINTER TERM 2023/24.

In addition, we are planning to re-design the AI course, to focus entirely on action-decision problems. In all likelihood, the winter edition 2023/24 will already have that new shape. So if you take the course again in 1.5 years, there will probably be many new things that you have not heard about yet. I apologize for the inconvenience, but this kind of thing cannot be avoided when adapting a course over time.

I wish you good luck in the upcoming exam phase,

best,

Jörg Hoffmann

 

 

===

Hi all,

another important thing to know for you in advance:

For logistical reasons (my research group will be doing Programmierung2 in summer 2023), the next edition of the AI course will be in the WINTER TERM 2023/24.

Please keep that in mind with a view on the amount of effort you invest in the current edition of the course; your next chance to pass the course will be in 1.5 years, not in 1 year.

best,

Jörg Hoffmann

Hi all,

as you can see from the timetable, I'll be giving the lectures in the next weeks, until June 9.

We'll still do the lecture online next week, May 9 and May 12, as I'd like the Covid numbers to go down a bit more.

However, after that, we'll go hybrid. Starting May 16, the lectures will be in GHH (Günter-Hotz Hörsaal); you will also be able to access them in zoom (same link as before); and they will also be streamed to YouTube (link posted in zoom chat, as before). I will try to watch the zoom chat during the lecture to answer questions from students participating online; but I'm not sure how well this will work out. If you want to participate actively, best come to GHH.

Note: The hybrid mode will be for the time being. It's not decided yet how we will continue after my part of the lecture.

Looking forward to see you in real life soon,

Jörg Hoffmann

 

Dear students,

We have now assigned the tutorials. If you log in to the CMS you can find your tutorial on the "personal status" page.

Forming Groups
You can now start/proceed forming groups. Please do so by using the forum. All members of a group need to be in the same tutorial (please compare the actual tutorial, not only the time slot, since there are tutorials using the same time slot in parallel). You do not need to mail the group members to anybody (the tutors or me). Please write them down on the solutions you hand in (as specified on the exercise sheets). You can start working in your new group right away.

Changing Tutorials
You can change your tutorial if and only if you find somebody who wants to swap the group with you. So if you want e.g. to be in Tutorial 3 and are in Tutorial 2, please go to the forum and find somebody who is in Tutorial 3 and wants to be in Tutorial 2. Then write a single e-mail per swap to me (Daniel Höller). Please use the following template and put the student you want to swap with in CC:

mail subject:
Changing Tutorials

mail content:
Full Name 1    from    Tutorial A    to    Tutorial B
Full Name 2    from    Tutorial B    to    Tutorial A

Best regards,
Daniel Höller

Dear students,

On your personal status page in CMS, you will now find the option to choose your tutorial preferences.

Please select your preferences till Tuesday, April 19, 15:00.

Further, we kindly ask those who have registered but have already decided not to attend the lecture to cancel their registration. This will make it much easier for us to plan the tutorials.

Best,
Daniel Höller

Dear students,

just to make this information very visible:

The lectures will be online in zoom first (for personal health reasons of the lecturers), and may move to presence lectures in Günter-Hotz-Hörsaal later. For up to date details, refer to the "Course Calendar". The link for the zoom lectures is https://cs-uni-saarland-de.zoom.us/j/98245956529?pwd=aU1oZ1ZUTGlxaEorUTNOR2FEU3Jhdz09

The online lectures will be streamed to YouTube (link provided in zoom chat at start of lecture), and links to the YouTube archive videos will be posted in CMS.

For presence teaching, we will look into hybrid teaching options. But we do not give a guarantee that all lectures will be available online.

In particular, the first lecture on Monday April 11, 10:15 -- 11:45, will be in zoom. I will explain organizational matters, and give a brief introduction to AI.

best,

Jörg Hoffmann