Automatically reducing use of cheat software in an online game environment

A computer-implemented method comprising:

• receiving, by one or more computing systems, training data about a plurality of decisions each specifying whether gameplay actions of a user in an interactive game environment result in access of the user to the interactive game environment being restricted due to use of cheat software to assist the gameplay actions;
• training, by the one or more computing systems, and based on the received training data, a deep neural network model for use in future classifying of further gameplay actions in the interactive game environment as resulting or not resulting in access to the interactive game environment being restricted due to use of cheat software to assist the further gameplay actions;
• using, by the one or more computing systems, the trained deep neural network model to control use of cheat software in the interactive game environment, including:
• determining, by the one or more computing systems, and for each of multiple users interacting with the interactive game environment, that the trained deep neural network model classifies observed gameplay actions of the user in the interactive game environment as unauthorized and resulting in access being restricted of the user to the interactive game environment due to the observed gameplay actions being similar to other user gameplay actions from the received data that were assisted by use of cheat software;
• obtaining, by the one or more computing systems, and for each of one or more first users of the multiple users, a verification that the observed gameplay actions of the first user are unauthorized and result in the access being restricted of the first user to the interactive game environment, including interacting with one or more other cheat software detection systems that are separate from the trained deep neural network model and that further assess the observed gameplay actions of the first user;
• and
• restricting, by the one or more computing systems, the access of each of the first users to the interactive game environment;
• and
• updating, by the one or more computing systems, the trained deep neural network model to improve further control of the use of cheat software in the interactive game environment, including:
• obtaining, by the one or more computing systems, and for each of one or more second users of the multiple users that are separate from the first users, a verification that the observed gameplay actions of the second user are not unauthorized and do not result in the access being restricted of the second user to the interactive game environment, including interacting with the one or more other cheat software detection systems that further assess the observed gameplay actions of the second user;
• and
• updating, by the one or more computing systems, the trained deep neural network model to reflect that, for each of the second users, the observed gameplay actions of the second user are not unauthorized and do not result in the access being restricted of the second user to the interactive game environment.