MeiChu Hackathon
2020 梅竹黑客松
Date
2020/10/24 (Sat.) - 2020/10/25 (Sun.)
Topic
Logitech International: Ethical gaming world
Motivation
There are increasingly more online games in recent years. When playing, we don’t want to encounter cheating behaviors that will decrease our will to play. Thus, we want to develop a cheating detection system that can detect unethical actions in online games.
Members
- Leo (Me)
- Bo-Wei Chen
- Jack Liu
- Bowen Lee
- Howard Wang
Ways of “Cheating”
- gaming hack
- AFK (away from keyboard)
- ask other people to play
- etc
Solution: Machine Learning
We surveyed a paper focusing on remote suspicious mouse intrusion, which gave us some inspiration. We believed that when a player cheats in a game, the mouse track must be different. So, we used random forest to train a one-class classifier, excepting the model after training could tell whether the input mouse track comes from the same person. By doing so, no matter the player installs gaming hack or asks his/her friends to play for him/her, our system could always tell us that the behavior was abnormal. Visit our presentation file for detailed descriptions.
Data Collection and Preprocessing
- Game: CS-GO
- Type: First-person shooter (FPS) game
- Data recording: Python
Awards
Out of hundreds of people, We won the FIRST place of Logitech and the FIRST place in the second round competing with other champions from other enterprises including LINE and Micron Taiwan. We were guaranteed to join the summer internship of Logitech and were reported in news.
Certificates of Merit
