A Persistent Spatial Semantic Representation for High-level Natural Language Instruction Execution
It addresses the challenge of bridging language and robot actions for non-experts in robotics, though it appears incremental as it builds on existing benchmarks and methods.
The paper tackles the problem of executing high-level natural language instructions for robotic tasks by proposing a persistent spatial semantic representation, achieving state-of-the-art results on the ALFRED benchmark without using step-by-step instructions.
Natural language provides an accessible and expressive interface to specify long-term tasks for robotic agents. However, non-experts are likely to specify such tasks with high-level instructions, which abstract over specific robot actions through several layers of abstraction. We propose that key to bridging this gap between language and robot actions over long execution horizons are persistent representations. We propose a persistent spatial semantic representation method, and show how it enables building an agent that performs hierarchical reasoning to effectively execute long-term tasks. We evaluate our approach on the ALFRED benchmark and achieve state-of-the-art results, despite completely avoiding the commonly used step-by-step instructions.