Awards

Abstract

We receive information about the world through our sensors and influence the world through our effectors. Such low-level experience has gradually come to play a greater role in AI over its 70-year history. I see this occurring in four steps, two of which are mostly past and two of which are in progress or yet to come. The first step was to view AI as the design of agents which interact with the world and thereby have experience; this viewpoint became prominent in the 1980s and 1990s (e.g., Russell & Norvig 1995). The second step was to view the goal of intelligence in terms of experience, as in the reward signal of optimal control and reinforcement learning. The reward formulation of goals is now widely used but rarely loved. Many would prefer to express goals in non-experiential terms, such as reaching a destination or benefiting humanity, but settle for reward because, as an experiential signal, reward is directly available to the agent without human assistance or interpretation. This is the pattern that we see in all four steps. Initially a non-experiential approach seems more intuitive, is preferred and tried, but ultimately proves a limitation on scaling; the experiential approach is more suited to learning and scaling with computational resources. The third step in the increasing role of experience in AI concerns the agent’s representation of the world’s state. Classically, the state of the world is represented in objective terms external to the agent, such as “the grass is wet” and “the car ten meters in front of me is slowing down”, or with probability distributions over world states such as in POMDPs and other Bayesian approaches. Alternatively, the state of the world can be represented experientially in terms of summaries of past experience (e.g., the last four Atari video frames input to DQN) or predictions of future experience (e.g., Predictive State Representations). The fourth step is potentially the biggest: world knowledge. Classically, world knowledge has always been expressed in terms far from experience, and this has limited its ability to be learned and maintained. Today we are seeing more calls for knowledge to be predictive and grounded in experience. After reviewing the history and prospects of the four steps, I propose a minimal architecture for an intelligent agent that is entirely grounded in experience.