Adaptive Skills, Adaptive Partitions (ASAP)
This work addresses the challenge of scaling up to lifelong learning agents by providing a general skill learning framework, though it appears incremental in building on existing skill learning concepts.
The paper tackles the problem of learning skills and where to apply them for lifelong learning agents, introducing the ASAP framework that can solve new tasks by adapting existing skills and requires less experience than learning from scratch.
We introduce the Adaptive Skills, Adaptive Partitions (ASAP) framework that (1) learns skills (i.e., temporally extended actions or options) as well as (2) where to apply them. We believe that both (1) and (2) are necessary for a truly general skill learning framework, which is a key building block needed to scale up to lifelong learning agents. The ASAP framework can also solve related new tasks simply by adapting where it applies its existing learned skills. We prove that ASAP converges to a local optimum under natural conditions. Finally, our experimental results, which include a RoboCup domain, demonstrate the ability of ASAP to learn where to reuse skills as well as solve multiple tasks with considerably less experience than solving each task from scratch.