[4C3-GS-10] AI application: prediction 2

[4C3-GS-10-05] Deck Generator in Game

〇Takayuki Shimotomai¹, Atsushi Takada¹, Naoki Hamada¹, Nyamkhuu Ganbat¹, Daichi Yamazaki¹ (1. KLab inc)

[[Online]]

Keywords:game, transformer, generative model

A lot of games for mobile devices have been developed and released. Among them, there is a type of game that players make a match by multiple cards called "deck". However, because it is difficult to design the game, developers need to adjust the balance of the game, and the costs are not low. To reduce the costs we have focused on deck auto-generation. In this study, we propose a deck generative model based on information about the opponent. The model is trained based on the user's game deck match data, and the resulting decks are verified and shown to be effective.