Minimum Description Length Control
This addresses the problem of efficient learning and adaptation across multiple tasks in reinforcement learning, though it appears incremental as it builds on existing principles like MDL and Bayesian inference.
The paper tackles multitask reinforcement learning by proposing a framework based on the minimum description length principle to learn common task structures, resulting in faster convergence and generalization to new tasks, with empirical effectiveness demonstrated on discrete and high-dimensional continuous control tasks.
We propose a novel framework for multitask reinforcement learning based on the minimum description length (MDL) principle. In this approach, which we term MDL-control (MDL-C), the agent learns the common structure among the tasks with which it is faced and then distills it into a simpler representation which facilitates faster convergence and generalization to new tasks. In doing so, MDL-C naturally balances adaptation to each task with epistemic uncertainty about the task distribution. We motivate MDL-C via formal connections between the MDL principle and Bayesian inference, derive theoretical performance guarantees, and demonstrate MDL-C's empirical effectiveness on both discrete and high-dimensional continuous control tasks.