Bayesian Relational Memory for Semantic Visual Navigation
Bayesian Relational Memory for Semantic Visual Navigation
复制标题
DOI：
10.1109/iccv.2019.00286
复制
发表时间：
2019-09
期刊：
影响因子：
--
通讯作者：
Yi Wu;Yuxin Wu;Aviv Tamar;Stuart J. Russell;Georgia Gkioxari;Yuandong Tian
中科院分区：
文献类型：
--
作者：
Yi Wu;Yuxin Wu;Aviv Tamar;Stuart J. Russell;Georgia Gkioxari;Yuandong Tian
我们引入了一种新的内存架构，贝叶斯关系内存（BRM），以提高泛化能力的语义视觉导航代理在看不见的环境中，代理被赋予一个语义目标导航。BRM采用语义实体上的概率关系图的形式（例如，房间类型），这允许（1）从训练环境中事先捕获布局，即，先验知识，（2）估计测试时的后验布局，即，存储器更新和（3）有效的导航规划。我们开发了一个BRM代理组成的BRM模块，用于产生子目标和控制的目标条件运动模块。当在看不见的环境中进行测试时，BRM代理的性能优于没有显式利用概率关系内存结构的基线。
We introduce a new memory architecture, Bayesian Relational Memory (BRM), to improve the generalization ability for semantic visual navigation agents in unseen environments, where an agent is given a semantic target to navigate towards. BRM takes the form of a probabilistic relation graph over semantic entities (e.g., room types), which allows (1) capturing the layout prior from training environments, i.e., prior knowledge, (2) estimating posterior layout at test time, i.e., memory update, and (3) efficient planning for navigation, altogether. We develop a BRM agent consisting of a BRM module for producing sub-goals and a goal-conditioned locomotion module for control. When testing in unseen environments, the BRM agent outperforms baselines that do not explicitly utilize the probabilistic relational memory structure.