Method and system for guaranteeing game quality by using artificial intelligence agent

1. A method of guaranteeing game quality by using an artificial intelligence (Al) agent, the method performed by a computer comprising:

extracting an inspection item list for inspecting quality of a target game;

extracting and storing log data corresponding to a test performance result for each item of the inspection item list;

performing imitation learning of an Al agent model on the basis of the stored log data;

performing an automatic test for inspecting quality of the target game by using the Al agent model on which the imitation learning is completed; and

automatically recording a bug and an error detected by the Al agent model,

wherein performing imitation learning of the Al agent model comprises performing imitation learning, which uses the stored log data as an input value of each of different Al agent models for each item of the inspection item list, and imitation learning which uses the stored log data as a random input value without classifying, for each item of the inspection item list, the stored log data by units of Al agent model, and

wherein performing the automatic test comprises performing the automatic test by using an Al agent model pool including an Al agent model for each item of the inspection item list on which the imitation learning is completed.

2. The method of claim 1 , wherein extracting and storing the log data comprises extracting and storing log data corresponding to a test performance result for each item of the inspection item list at least once or more, on the basis of a requirement method of an AI agent model which is a learning target.

3. The method of claim 1 , wherein extracting and storing the log data comprises extracting and storing, as log data of each item of the inspection item list, in-game action and environment interaction information about the target game executed by a manager.

4. A system for guaranteeing game quality by using an artificial intelligence (Al) agent, the system comprising:

a memory storing a program for automatically detecting a bug and an error of a target game on the basis of an Al agent model; and

a processor configured to execute the program stored in the memory,

wherein, by executing the program, the processor:

receives a test performance result for each item of an inspection item list for inspecting quality of the target game,

extracts and stores log data corresponding to a test performance result for each item of the inspection item list of the target game,

performs imitation learning of the Al agent model on the basis of the stored log data, performs an automatic test for inspecting quality of the target game by using the Al agent model on which the imitation learning is completed, and

automatically records a bug and an error detected by the Al agent model,

wherein the processor performs imitation learning, which uses the stored log data as an input value of each of different Al agent models for each item of the inspection item list, and imitation learning which uses the stored log data as a random input value without classifying, for each item of the inspection item list, the stored log data by units of Al agent model, and

wherein the processor performs an automatic test for inspecting quality of the target game by using an Al agent model pool including an Al agent model for each item of the inspection item list on which the imitation learning is completed.

5. The system of claim 4 , wherein the processor extracts and stores log data corresponding to a test performance result for each item of the inspection item list at least once or more, on the basis of a requirement method of a learning target model.

6. The system of claim 4 , wherein the processor extracts and stores, as log data of each item of the inspection item list, in-game action and environment interaction information about the target game executed by a manager.