Growing Q-Networks: Solving Continuous Control Tasks with Adaptive Control Resolution

arXiv:2404.04253v1 Announce Type: new
Abstract: Recent reinforcement learning approaches have shown surprisingly strong capabilities of bang-bang policies for solving continuous control benchmarks. The underlying coarse action space discretizations often yield favourable exploration characteristics while final performance does not visibly suffer in the absence of action penalization in line with optimal control theory. In robotics applications, smooth control signals are commonly preferred to reduce system wear and energy efficiency, but action costs can be detrimental to exploration during early training. In …

abstract arxiv benchmarks capabilities continuous control cs.ai cs.lg cs.ro exploration line networks performance policies reinforcement reinforcement learning resolution space tasks type