当前位置: X-MOL 学术arXiv.cs.CV › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Utilising Prior Knowledge for Visual Navigation: Distil and Adapt
arXiv - CS - Computer Vision and Pattern Recognition Pub Date : 2020-04-07 , DOI: arxiv-2004.03222
M. Mahdi Kazemi Moghaddam, Qi Wu, Ehsan Abbasnejad and Javen Shi

We, as humans, can impeccably navigate to localise a target object, even in an unseen environment. We argue that this impressive ability is largely due to incorporation of \emph{prior knowledge} (or experience) and \emph{visual cues}--that current visual navigation approaches lack. In this paper, we propose to use externally learned prior knowledge of object relations, which is integrated to our model via constructing a neural graph. To combine appropriate assessment of the states and the prior (knowledge), we propose to decompose the value function in the actor-critic reinforcement learning algorithm and incorporate the prior in the critic in a novel way that reduces the model complexity and improves model generalisation. Our approach outperforms the current state-of-the-art in AI2THOR visual navigation dataset.

中文翻译:

利用视觉导航的先验知识:提炼和适应

作为人类,即使在看不见的环境中,我们也可以无可挑剔地导航以定位目标对象。我们认为这种令人印象深刻的能力主要是由于\emph{先验知识}(或经验)和\emph{视觉线索}的结合——目前的视觉导航方法缺乏。在本文中,我们建议使用外部学习的对象关系先验知识,通过构建神经图将其集成到我们的模型中。为了结合对状态的适当评估和先验(知识),我们建议分解 actor-critic 强化学习算法中的价值函数,并以一种降低模型复杂性并提高模型泛化的新方式将先验合并到 critic 中。我们的方法在 AI2THOR 视觉导航数据集中优于当前最先进的方法。
更新日期:2020-04-08
down
wechat
bug