Metric Learning for Generalizing Spatial Relations to New Objects

TL;DR

This paper presents a metric learning approach that enables robots to generalize spatial relations to new objects by reasoning about relation similarities, facilitating interactive learning from few examples.

Contribution

It introduces a novel distance metric learning method for spatial relation generalization, allowing robots to learn arbitrary relations from limited, non-expert demonstrations.

Findings

Effective reasoning about a spectrum of spatial relations

Successful generalization to new objects in real-world scenarios

Enables interactive, few-shot learning of spatial relations

Abstract

Human-centered environments are rich with a wide variety of spatial relations between everyday objects. For autonomous robots to operate effectively in such environments, they should be able to reason about these relations and generalize them to objects with different shapes and sizes. For example, having learned to place a toy inside a basket, a robot should be able to generalize this concept using a spoon and a cup. This requires a robot to have the flexibility to learn arbitrary relations in a lifelong manner, making it challenging for an expert to pre-program it with sufficient knowledge to do so beforehand. In this paper, we address the problem of learning spatial relations by introducing a novel method from the perspective of distance metric learning. Our approach enables a robot to reason about the similarity between pairwise spatial relations, thereby enabling it to use its previous knowledge when presented with a new relation to imitate. We show how this makes it possible to learn arbitrary spatial relations from non-expert users using a small number of examples and in an interactive manner. Our extensive evaluation with real-world data demonstrates the effectiveness of our method in reasoning about a continuous spectrum of spatial relations and generalizing them to new objects.