Processant ...

Chess playing engines such as AlphaZero have delivered phenomenal improvements to game-playing strategies using reinforcement learning [1, 2]. Frameworks like the universal chess interface assist these engines in communicating with user interfaces to analyze chess variant positions and to strategize the next moves by applying discrete probabilities. The objective of this project is to apply game-theoretic AI to test security protocols. In this work, the security protocol [3] becomes the vocabulary of the gameplay, and the reinforcement learning engine is used to exhaustively explore and evaluate all gameplay variants with a view to identify and detect errors in the protocol. 1. Ari Juels and Madhu Sudan. A fuzzy vault scheme. Designs, Codes and Cryptography, 38(2):237–257, Feb 2006. 2. David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, Timothy P. Lillicrap, Karen Simonyan, and Demis Hassabis. Mastering chess and shogi by self-play with a general reinforcement learning algorithm. CoRR, abs/1712.01815, 2017. 3. Rolando Trujillo-Rasua, Benjamin Martin, and Gildas Avoine. The poulidor distance-bounding protocol. In Siddika Berna Ors Yalcin, editor, Radio Frequency Identification: Security and Privacy Issues, pages 239–257, Berlin, Heidelberg, 2010. Springer Berlin Heidelberg.