Statistics Lab

Dear all,

we would like to share some thought about ALA with you and thank everybody who gave feedback until now.
Regards, Verena Wolf.

Why practice with AI?

Like it or not, working with AI is becoming part of most jobs. It makes sense to start learning how to use it well now rather than later. Part of that skill is understanding where AI is actually good and where it produces nonsense that needs to be critically checked. A good AI tutor can also explain things in different ways, give you instant feedback, and help you when you're stuck — without just giving you the answer.

Why as a research project?

ALA is a research project at DFKI and Saarland University. We're studying how to build AI learning tools that actually help you think, not think for you. By using ALA in the Statistics Lab, you get a useful tool and we learn how to make it better. Win-win.

Why is it mandatory?

Same reason homework is mandatory in introductory courses: regular engagement with the material is a prerequisite for exam admission. ALA exercises are just another form of that.

Why pass/fail instead of partial credit?

Each ALA exercise is graded pass/fail. Since the goal is to encourage engagement with the material and AI rather than evaluate every detail for correctness, partial credit is not meaningful or appropriate in this context. The points per exercise vary and are scaled to match the other coursework components.

IMPORTANT

If something doesn't work, if a conversation is unhelpful or confusing, please use the thumbs-up/thumbs-down buttons. We're improving the system continuously and your feedback directly helps.

Dear students,

ALA Submission Guidelines

As there seems to be some confusion regarding ALA and what constitutes a valid submission, we would like to clarify this.

We aim to encourage healthy and meaningful use of LLMs and chatbots. ALA was designed with this goal in mind. To assess whether you have actively worked with ALA, we are interested in seeing your thought process and a clear progression toward solving the exercise.

What we are looking for:

• your thought process
• whether the chat meaningfully covers the exercise
• evidence that you worked toward a solution
• your own ideas and intermediate steps

What we are not looking for:

• a perfectly correct final solution
• whether ALA flagged your submission as correct
• perfectly formatted or error-free LaTeX

Regarding LaTeX:

You are not required to use LaTeX to solve the exercises. Plain text, informal notation, or partially correct mathematical expressions are completely fine as long as your reasoning is understandable. Modern LLMs are generally able to interpret your intent, and this is sufficient for us as well.

What will be rejected (non-exhaustive list):

• chats that only present a final solution, even if you arrived at it on your own
• chats in a language other than English
• chats that are unrelated to the exercise
• chats where only hints are requested but no own reasoning or follow-up work is shown
• chats that do not demonstrate any meaningful engagement with the task

If in doubt, ask in the forum. Our tutors will happily help clarify any remaining questions.

Please also note that ALA chat submissions are mandatory for exam admission. You must participate in 9 out of 11 units, including both the exercise sheet and the ALA submission. Partial participation (only the exercise sheet or only the ALA chat) is not sufficient.

---

Points & Status Updates

Points have now been published for:

• Mini test unit 1 & 2
• Exercise sheet unit 2
• ALA unit 1 (test run)

Due to illness, not all points have been entered yet. These will be added soon, and your personal status page will update automatically.

For all students, a "passed" or "not passed" status is now shown for unit 1, depending on whether you participated in tutorial 1. For the remaining units, this status will be displayed once the results for the exercise sheet and the ALA submission of the respective unit have been published.

Best regards,
Your StatLab team

Dear Students,

The second exercise sheet has just been released. With that, ALA has officially started.

Since there has been some confusion regarding the exam admission:

Just like the tutorial exercise sheets and mini-tests, submitting ALA dialogs is mandatory to qualify for the exam.

You are allowed to miss at most one mini-test and two units without providing a reason; keep in mind that missing an ALA dialog or

exercise sheet submission constitutes a missed unit. Please note that you cannot take a mini-test or submit an exercise sheet in a

tutorial group other than your assigned one. If you are unsure whether you are still allowed to miss a mini-test or unit, please contact

us well in advance. In case of illness, you may make up for a missed mini-test or exercise sheet if you can provide a doctor's note.

In that case, please contact our office at: nmm-office@dfki.de

Lastly, we have also published the solutions to last week's exercise sheet.

All the best

Your Statlab Team

Dear students,

a video showcasing the use of ALA is now available under Information --> Materials --> 0 ALA Showcase, or directly via the following link: https://cms.sic.saarland/statlab_26/dl/10/0_ALA_Showcase.mp4

Please make sure to familiarize yourselves with this tool, as the first mandatory ALA exercises will begin next week.

If you encounter any issues or have questions, please contact us via the Forum so we can help you as quickly as possible.

Best regards,
Your StatLab Team

Dear Students,

The first exercise sheet has just been released. You can find it under Information --> Materials.

With the release of the exercise sheet, we have also granted you access to ALA: https://ala.sb.dfki.de/

A video demo showcasing how ALA works and what makes a good/bad dialog will be released in the coming days. You can use the time before the video comes out to experiment with ALA yourselves.

However, please note that you can submit each exercise only once. Hence, if you are unsure if your dialog is acceptable, we strongly discourage you from submitting it before the video comes out.

All the best,

Your Statlab Team

Dear Students,

This week, we will be conducting an ALA test run. Participation is not mandatory, and if you choose not to participate, this week will not be factored into your exam admission calculation.

However, this is an opportunity to receive feedback on what constitutes a good or bad dialogue. As an added bonus, if your score on this test run improves your overall points, we will include it in your exam admission calculation. Therefore, participation is still strongly recommended.

You will receive feedback no later than Friday, April 24. Please note that by this time, the 2nd ALA week will have already started.

All the best,

Your Statlab Team

Dear students,

We have published the tutorial slot assignments. You can check your assigned slot on your "Personal Status" page. Please contact us immediately if you encounter any issues.

Also note that the first tutorials will take place tomorrow and the day after (April 13th & 14th), respectively.

Best regards,
Your StatLab Team

Dear participants,

Tomorrow at 8:30, we will have our second on-site meeting in the GHH lecture hall and discuss the contents of the first Unit based on several examples. 
Online participation is possible via Teams: 
GHH Tutorial | Besprechung beitreten | Microsoft Teams
 
Kind regards,
V. Wolf.

Dear students,

Please remember to register for this course til Thursday, April 9 at 23:59.

Moreover, registration for the tutorials is now open. Please submit your preferences under the Personal Status section.

You can update your preferences until Friday, April 10 at 23:59.

Additionally, materials for the first unit have already been posted under Information --> Materials.

There you can also find the slides from today's lecture covering the course organization.

All the best,

Statlab Team