Justin Wood, Indiana University
Reverse Engineering the Origins of Visual Intelligence
How do brains produce intelligent behavior, and how can we replicate intelligence in machines? To address these questions, my lab uses a two-pronged approach. First, we perform automated controlled-rearing experiments, using newborn chicks as a model system. We raise chicks in strictly controlled virtual worlds and record their behavior 24/7 as they learn to perceive and understand their environment. Fueled by interactive virtual reality chambers, we can explore how foundational abilities (e.g., object recognition, intuitive physics) emerge in newborn brains. Second, we perform parallel experiments on autonomous artificial agents, using virtual controlled-rearing chambers. We raise newborn animals and artificial agents in the same environments, and test whether they develop the same abilities when given the same experiences. The agents’ brains can be equipped with different biologically-inspired learning mechanisms (e.g., deep reinforcement learning, curiosity-driven learning), so by comparing the animals and agents, we can test which learning mechanisms are needed to mimic the development of visual intelligence. Our ultimate goal is to reverse engineer the origins of intelligence and build machines that learn like newborn brains.