arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1807.06583cs.CVcs.RO
收藏

示教学习的可解释潜在空间

Interpretable Latent Spaces for Learning from Demonstration

University of Edinburgh(爱丁堡大学)

Yordan Hristov, Alex Lascarides, Subramanian Ramamoorthy

摘要:

有效的人机交互,例如在从人类演示的机器人学习中,需要学习代理能够将抽象概念(例如包含在指令中的那些)基于来自世界的相应高维感觉输入流。诸如深度神经网络的模型通过其大参数空间具有高容量,可以用于将高维传感数据压缩为较低维表示。这些低维表示有助于符号基础,但可能无法保证表示是人类可解释的。我们提出了一种方法,它利用用户定义的符号和它们相应的感官观察分组,以对齐学习压缩的潜在表示与抽象标签中包含的语义概念。我们通过模拟和真实世界的对象数据的实验证明了这一点,表明这种对齐可以在物理符号接地的过程中实现。

英文摘要:

Effective human-robot interaction, such as in robot learning from human demonstration, requires the learning agent to be able to ground abstract concepts (such as those contained within instructions) in a corresponding high-dimensional sensory input stream from the world. Models such as deep neural networks, with high capacity through their large parameter spaces, can be used to compress the high-dimensional sensory data to lower dimensional representations. These low-dimensional representations facilitate symbol grounding, but may not guarantee that the representation would be human-interpretable. We propose a method which utilises the grouping of user-defined symbols and their corresponding sensory observations in order to align the learnt compressed latent representation with the semantic notions contained in the abstract labels. We demonstrate this through experiments with both simulated and real-world object data, showing that such alignment can be achieved in a process of physical symbol grounding.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑