Learning Lexical Entries for Robotic Commands using Crowdsourcing

TL;DR

This paper explores translating natural language robotic commands into a robot-specific language using crowdsourcing and generative models to improve robot understanding of diverse spatial instructions.

Contribution

It introduces a crowdsourcing approach to collect commands and employs a generative translation model to convert natural language into robot language, addressing linguistic variability.

Findings

Crowdsourcing effectively gathers diverse robotic commands.

Generative translation models can map natural language to robot language.

The approach simulates human-robot interaction for better understanding.

Abstract

Robotic commands in natural language usually contain various spatial descriptions that are semantically similar but syntactically different. Mapping such syntactic variants into semantic concepts that can be understood by robots is challenging due to the high flexibility of natural language expressions. To tackle this problem, we collect robotic commands for navigation and manipulation tasks using crowdsourcing. We further define a robot language and use a generative machine translation model to translate robotic commands from natural language to robot language. The main purpose of this paper is to simulate the interaction process between human and robots using crowdsourcing platforms, and investigate the possibility of translating natural language to robot language with paraphrases.