We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Robotics

Title: A Persistent Spatial Semantic Representation for High-level Natural Language Instruction Execution

Abstract: 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.
Comments: Presented at CoRL 2021
Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as: arXiv:2107.05612 [cs.RO]
  (or arXiv:2107.05612v3 [cs.RO] for this version)

Submission history

From: Valts Blukis [view email]
[v1] Mon, 12 Jul 2021 17:47:19 GMT (2414kb,D)
[v2] Tue, 19 Oct 2021 13:29:11 GMT (2420kb,D)
[v3] Sun, 28 Nov 2021 19:33:23 GMT (2421kb,D)

Link back to: arXiv, form interface, contact.