Temporally Abstract Partial Models
This work addresses a specific challenge in reinforcement learning for improving agent efficiency, but it is incremental as it builds on existing option models.
The paper tackles the problem of enabling reinforcement learning agents to reason about temporally abstract actions (options) that are only feasible in certain situations, by defining affordances for options and developing partial option models. It demonstrates empirically that these models can improve planning efficiency.
Humans and animals have the ability to reason and make predictions about different courses of action at many time scales. In reinforcement learning, option models (Sutton, Precup \& Singh, 1999; Precup, 2000) provide the framework for this kind of temporally abstract prediction and reasoning. Natural intelligent agents are also able to focus their attention on courses of action that are relevant or feasible in a given situation, sometimes termed affordable actions. In this paper, we define a notion of affordances for options, and develop temporally abstract partial option models, that take into account the fact that an option might be affordable only in certain situations. We analyze the trade-offs between estimation and approximation error in planning and learning when using such models, and identify some interesting special cases. Additionally, we demonstrate empirically the potential impact of partial option models on the efficiency of planning.