Hierarchical Primitive Composition: Simultaneous Activation of Skills With Inconsistent Action Dimensions in Multiple Hierarchies

Deep reinforcement learning has shown its effectiveness in various applications, providing a promising direction for solving tasks with high complexity. However, naively applying classical RL complex long-horizon task single control policy is inefficient. Thus, modularization tackles this problem by set of modules that are mapped to primitives and properly orchestrating them. In study, we further expand the discussion incorporating simultaneous activation skills structuring them into multiple hierarchies recursive fashion. Moreover, sought devise an algorithm can orchestrate different action spaces via multiplicative Gaussian distributions, which highly increases reusability. By exploiting modularity, interpretability also be achieved observing used new if each known. We demonstrate how proposed scheme employed practice pick place 6 DoF manipulator, examine effects property from ablation studies.