Lecture
This lecture focuses on perceptual modeling and spatial thinking within the context of visual intelligence. It begins with a recap of previous theories, including structuralism and Gestaltism, emphasizing how these theories contribute to understanding visual perception. The instructor discusses the concept of vision as a reconstruction of reality, highlighting the importance of both bottom-up and top-down processing in interpreting visual stimuli. The evolutionary utility of vision is explored, illustrating how observers construct mental models of their environments based on sensory input. The lecture also delves into visual illusions, demonstrating how these can reveal inaccuracies in our mental models. The concept of cognitive maps is introduced, showcasing experiments that illustrate how both humans and animals navigate and understand their environments. The discussion extends to model-based reinforcement learning and the significance of spatial organization in cognitive processes, culminating in a comprehensive understanding of how perception and spatial thinking are intertwined in visual intelligence.