AI Bot Trainer and Data Labeler for Game AI Project
Designed and trained an AI bot to play a Street Fighter-style game using a Multi-Layer Perceptron model. Manually performed a variety of in-game moves to collect gameplay data, then applied labeling to this data for training. Utilized labeled datasets to improve the model's predictive decision-making ability in-game. • Collected and labeled gameplay data through manual actions. • Prepared training datasets focused on move outcomes and scenarios. • Applied custom labels to support model fine-tuning for optimal bot performance. • Enhanced model accuracy and in-game intelligence through targeted data labeling.