These are general rules for IOAI selections, which may be updated before the competition or have specific differences for certain days or tasks. These updates will be communicated to selection participants in emails/Discord messages, so please read communication emails carefully.

1. **By default, the use of pre-trained models and additional data is not allowed**, unless otherwise specified in the task statement.

2. **Participants must solve these competitions solely on their own and without any external help** (e.g., asking teachers, friends or sharing data/ideas with other participants is prohibited).

3. **Use of LLM chats, ChatGPT, Copilot or similar AI tools**, including using LLM APIs, is prohibited, except the inline autocompletions on https://colab.research.google.com/

4. **On the leaderboard of every competition, name yourself as "[City] Name Surname"** (e.g., "[Paphos] Alexander Avdiushenko").

5. **You are only allowed to use Kaggle website and Google Colab:** the pages of selection competitions and Kaggle Notebooks.
   This also means that the use of VSCode, PyCharm is prohibited; using online documentation for ML libraries is allowed.

6. **You cannot use multiple Kaggle accounts during selections.**
   If you have multiple accounts and your computational quota is exhausted, selecting an account that has it fresh and full is allowed.
   If you use different accounts for different tasks, we may not join your data correctly and sum up your points correctly, so this is prohibited.

7. **You must disconnect Google Cloud APIs, Google SDK or any other additional things** connected to your Kaggle account that could grant additional quota, resources or anything giving an unfair advantage compared to others who do not have it in Kaggle Notebooks.

8. **Contestants cannot use cell phones or other devices during the competition.**

9. **Contestants are prohibited from engaging in any form of cheating**, including tampering with the contest system, direct or indirect communication with other contestants or people outside the contest hall during the contest, bringing prohibited items into the contest hall, or attempting unauthorized access to test data.

10. **The screen of each contestant’s computer may be recorded or remotely proctored.**

11. **Violating any of these rules may result in disqualification or other penalties.**

## Scoring for ML tasks

The maximum score for each task is **100 points**. The baseline score is **0 points**. Both are calculated on Private LB.

More precisely: for each task, the score of a submission is normalized using the following formula:

```
Norm_Score = (Submission_Score - Min_Score) / (Max_Score - Min_Score) * 100
```

* **Submission_Score** = The raw score obtained for the submission (e.g., accuracy, AUC, F1-score, or any other metric specified in the task statement).
* **Min_Score** = The score of the baseline solution.
* **Max_Score** = The highest unnormalized score achieved by any contestant’s submission during the contest.

---

## And as always – for everyone:

* Don't cheat!
* Apply yourself!
* Have fun!