deep reinforcement learning

Murray Shanahan explains how his work moved from logic programming and classical AI through computational neuroscience to machine learning and deep reinforcement learning. He uses the frame problem to show why intelligent systems need to focus on what matters in a situation, and discusses how symbolic AI ideas might help current learning systems. Reflecting on Ex Machina, he describes advising Alex Garland, the film’s portrayal of embodiment and consciousness, and the hidden code Easter egg he wrote for the film. He also considers AI’s distance from human-level intelligence, the limits of Asimov’s laws as engineering rules, and how technology can augment human abilities.