[MAX] ML Anti XRay

AI Anti-Xray Detection Plugin

This plugin is designed to help Minecraft server administrators detect and prevent x-ray cheating using a simple AI model. It monitors players' ore mining behavior and collects various data points to determine if a player is likely using x-ray mods.

---

Key Features

• AI-Driven Detection:
  The plugin uses two sets of model weights (for legit and x-ray behavior) to compute a confidence score indicating the likelihood that a player is x-raying.

• Comprehensive Data Collection:

  • Distance Calculation: Tracks the distance from the last mined ore.
  • Exposure Analysis: Determines whether an ore was naturally visible or revealed due to nearby block breaking.
  • Ore Vein Grouping: Data is only recorded once per ore vein to avoid duplicate counts.
  • Mining History: Maintains a rolling history of ore mining events (default: 10 minutes).
  • Rotation from last ore: Checks how big of a rotation you'd have to do in order to reach an ore from the last one.

• Dynamic Training System:
  Use the /train <legit|xray> command to adjust the AI model weights based on player behavior. This allows the model to learn and improve over time:

  • Legit Training: Adjusts the model if a player is confidently a legitimate miner.
  • Xray Training: Adjusts the model when a player is detected as using x-ray mods.

• Customizable Configuration:
  Configure the plugin via config.yml to set:

  • Flag threshold (confidence percentage).
  • Flag messages.
  • List of ores to monitor.
  • Mining time window for data collection.
  • And more...

• Automatic File Management:
  The plugin automatically creates necessary folders and files (e.g., ML/legit.dat, ML/xray.dat) if they do not exist, ensuring smooth first-time setup.

---

How It Works

1. Data Collection:

   • When a player breaks a block, the plugin checks if it's an ore of interest.
   • If a non-ore block is broken, the plugin scans adjacent blocks for ores and marks them as "revealed."
   • When an ore is mined, the plugin checks if it was naturally exposed (i.e., visible without needing to break any adjacent block) before recording the event.

2. Feature Extraction:
   The plugin calculates:

   • Player rotation difference to the ore.
   • Distance from the previous ore break.
   • Exposure status (using the revealed ore list).
   • Count of similar ores mined within the configured time window.

3. AI Prediction:
   The collected features are processed through a weighted model, and a confidence score is computed that indicates the likelihood of x-ray usage.

4. Training & Adaptation:
   The /train command allows administrators or trusted players to train the model based on observed behavior, updating the weights stored in ML/legit.dat or ML/xray.dat.

5. Flagging Suspicious Behavior:
   If the confidence score exceeds the configured threshold, the plugin flags the player which then notifies staff of their potentially unfair behaviour.

---

Usage

• Installation:
  Place the plugin jar into your server's plugins folder. On first run, the necessary configuration and ML data files will be created automatically.

• Configuration:
  Edit config.yml in the plugin’s data folder to customize the detection thresholds, messages, and ore list.

• Commands:

  • /train <legit|xray>
    Use this command to train the AI model based on a player's recent mining behavior.

---

Conclusion

This AI Anti-Xray Detection Plugin leverages in-game data to help maintain a fair play environment on your server by dynamically learning and adapting to player behavior. Enjoy a smarter, self-improving system for combatting x-ray cheats!

Feel free to reach out with feedback or suggestions to improve the plugin further. Happy mining!

https://bstats.org/plugin/bukkit/MLAntiXray/25104