Dec, 2022

学习图搜索启发式

TL;DRPHIL (Path Heuristic with Imitation Learning) is a neural architecture and training algorithm that uses imitation learning and graph representation learning to discover graph search and navigation heuristics from data, reducing the number of explored nodes by 58.5% compared to state-of-the-art methods on benchmark datasets and allowing for fast planning in time-critical robotics domains.