We present ReViND -- a method that combines the strength of offline RL with topological graphs to get customizable long-range navigation behaviors.
"Offline Reinforcement Learning with Visual Navigation"
Dhruv Shah*, Arjun Bhorkar*, Hrish Leen, Ilya Kostrikov, Nick Rhinehart, Sergey Levine
Berkeley AI Research (UC Berkeley)
Presented at Conference on Robot Learning (CoRL) 2022, Auckland NZ
Project Page: [ Ссылка ]
PDF: [ Ссылка ]_
Timeline:
-------------
00:00 Introduction
00:15 Motivation
00:51 Reinforcement Learning
01:42 Topological Graphs
02:43 Offline Values as Distances
03:40 System Overview
04:22 Results
