Reinforcement Learning (Atari Games)

Dataset: The experiments are conducted using the Seaquest environment from the OpenAI Gym framework, which simulates the Atari 2600 game Seaquest. The dataset consists of RGB frames (210x160x3) generated dynamically during training. These frames are preprocessed by converting to grayscale, resizing to 84x84 pixels, and stacking four consecutive frames to form a 4x84x84 tensor, capturing temporal dynamics of the game state. No external or pre-collected dataset is used; the data is produced through real-time interaction with the Gym environment.