This is a Python implementation of the papers: [1] [2]
The framework in the paper predicts cheats in computer games leveraging machine learning models over captured encrypted game traffic data during game play. We proposed an efficient transfer learning approach to address limited labeled data problems for training models.
This python implementation of the transfer learning approach extends RuLSIF by automatically learning the α-relative density ratio parameter for superior relative density ratio estimation.
More on RuLSIF: https://github.com/BoDong111/Multistream-Classification
[1] Bo Dong, Md Shihabul Islam, Swarup Chandra, Latifur Khan, and Bhavani Thuraisingham. "GCI: A transfer learning approach for detecting cheats of computer game." In 2018 IEEE International Conference on Big Data (Big Data), pp. 1188-1197. IEEE, 2018.
[2] Md Shihabul Islam, Bo Dong, Swarup Chandra, Latifur Khan, and Bhavani Thuraisingham. "GCI: A GPU-Based Transfer Learning Approach for Detecting Cheats of Computer Game." IEEE Transactions on Dependable and Secure Computing 19, no. 2 (2020): 804-816.