BriefGPT.xyz
Apr, 2022
带有学习路径排名的粗到细的Q-attention
Coarse-to-Fine Q-attention with Learned Path Ranking
HTML
PDF
Stephen James, Pieter Abbeel
TL;DR
提出了 Learned Path Ranking 方法，通过对路径生成方法（包括路径规划、贝塞尔曲线采样和学习策略）的排序学习，可以在保持以前样本效率的同时，实现更多的任务，并在16个 RLBench 任务中进行了基准测试。
Abstract
We propose
learned path ranking
(LPR), a method that accepts an end-effector goal pose, and learns to rank a set of goal-reaching paths generated from an array of
path generating methods
, including: path planning
→