English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This letter introduces ERRA, an embodied learning architecture that enablesrobots to jointly obtain three fundamental capabilities (reasoning, planning,and interaction) for solving long-horizon language-conditioned manipulationtasks. ERRA is based on tightly-coupled probabilistic inferences at twogranularity levels. Coarse-resolution inference is formulated as sequencegeneration through a large language model, which infers action language fromnatural language instruction and environment state. The robot then zooms to thefine-resolution inference part to perform the concrete action corresponding tothe action language. Fine-resolution inference is constructed as a Markovdecision process, which takes action language and environmental sensing asobservations and outputs the action. The results of action execution inenvironments provide feedback for subsequent coarse-resolution reasoning. Suchcoarse-to-fine inference allows the robot to decompose and achieve long-horizontasks interactively. In extensive experiments, we show that ERRA can completevarious long-horizon manipulation tasks specified by abstract languageinstructions. We also demonstrate successful generalization to the novel butsimilar natural language instructions.

ERRA: An Embodied Representation and Reasoning Architecture for Long-horizon Language-conditioned Manipulation Tasks