arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1810.01344cs.LGstat.ML
收藏

感觉运动预测中空间结构的无监督涌现

Unsupervised Emergence of Spatial Structure from Sensorimotor Prediction

SoftBank Robotics Europe(软银机器人欧洲公司)

Alban Laflaquière, Michael Garcia Ortiz

摘要:

尽管它在机器人应用中无处不在,空间知识的性质和机制,其出现在自主智能体仍然知之甚少。最近的理论工作表明,空间的概念可以通过捕获代理的原始感觉运动经验中的空间结构引起的不变量来接地。此外,据推测,捕捉这些不变量是有益的幼稚代理试图预测其感觉运动的经验。在某些探索性条件下,空间表征应该是学习预测的副产品。我们提出了一个简单的感觉运动预测方案,将其应用于不同的代理和类型的探索,并评估这一假设的相关性。我们表明,一个天真的代理可以捕捉到的拓扑结构和度量规则的空间配置没有任何先验知识,也没有多余的监督。

英文摘要:

Despite its omnipresence in robotics application, the nature of spatial knowledge and the mechanisms that underlie its emergence in autonomous agents are still poorly understood. Recent theoretical work suggests that the concept of space can be grounded by capturing invariants induced by the structure of space in an agent's raw sensorimotor experience. Moreover, it is hypothesized that capturing these invariants is beneficial for a naive agent trying to predict its sensorimotor experience. Under certain exploratory conditions, spatial representations should thus emerge as a byproduct of learning to predict. We propose a simple sensorimotor predictive scheme, apply it to different agents and types of exploration, and evaluate the pertinence of this hypothesis. We show that a naive agent can capture the topology and metric regularity of its spatial configuration without any a priori knowledge, nor extraneous supervision.

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。

↑