Intrinsically Motivated Learning of Causal World Models
This work addresses the challenge of achieving more general intelligence in AI systems, which is a foundational problem for the field, though it appears incremental as it builds on existing ideas in world modeling and causal inference.
The paper tackles the problem of limited skill transfer and generalization in AI by proposing to build causal world models through intrinsically motivated actions that collect interventional data, aiming to improve the inference of environmental causal structures.
Despite the recent progress in deep learning and reinforcement learning, transfer and generalization of skills learned on specific tasks is very limited compared to human (or animal) intelligence. The lifelong, incremental building of common sense knowledge might be a necessary component on the way to achieve more general intelligence. A promising direction is to build world models capturing the true physical mechanisms hidden behind the sensorimotor interaction with the environment. Here we explore the idea that inferring the causal structure of the environment could benefit from well-chosen actions as means to collect relevant interventional data.