Though computational cognitive architectures have been used to study several processes associated with human behavior, the study of integration of affect and emotion in these processes has been relatively sparse. Theory from affective science and affective neuroscience can be used to systematically integrate affect into cognitive architectures, particularly in areas where cognitive system behavior is known to be associated with physiological structure and behavior. I introduce a unified theory and model of human behavior that integrates physiology and primal affect with cognitive processes in a cognitive architecture. This new architecture gives a more tractable, mechanistic way to simulate affect-cognition interactions to provide specific, quantitative predictions. It considers affect as a lower-level, functional process that interacts with cognitive processes (e.g., declarative memory) to result in emotional behavior. This formulation makes it more straightforward to connect these affective representations with other related moderating processes that may not specifically be considered as emotional (e.g., thirst or stress). An improved understanding of the architecture that constrains our behavior gives us a better opportunity to comprehend why we behave the way we do and how we can use this knowledge to recognize and construct a more ideal internal and external environment.

Teaching a computer to play video games has generally been seen as a reasonable benchmark for developing new AI techniques. In recent years, extensive research has been completed to develop reinforcement learning (RL) algorithms to play various Atari 2600 games, resulting in new applications of algorithms such as Deep Q-Learning or Policy Gradient that outperform humans. However, games from Super Nintendo Entertainment System (SNES) are far more complicated than Atari 2600 games as many of these state-of-the-art algorithms still struggle to perform on this platform. In this paper, we present a new platform to research algorithms on SNES games and investigate NeuroEvolution of Augmenting Topologies (NEAT) as a possible approach to develop algorithms that outperform humans in SNES games.