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

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

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

Computer Science > Computation and Language

Title: Continual Learning for Grounded Instruction Generation by Observing Human Following Behavior

Abstract: We study continual learning for natural language instruction generation, by observing human users' instruction execution. We focus on a collaborative scenario, where the system both acts and delegates tasks to human users using natural language. We compare user execution of generated instructions to the original system intent as an indication to the system's success communicating its intent. We show how to use this signal to improve the system's ability to generate instructions via contextual bandit learning. In interaction with real users, our system demonstrates dramatic improvements in its ability to generate language over time.
Comments: To appear in TACL 2021. The arXiv version is a pre-MIT Press publication version
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2108.04812 [cs.CL]
  (or arXiv:2108.04812v1 [cs.CL] for this version)

Submission history

From: Noriyuki Kojima [view email]
[v1] Tue, 10 Aug 2021 17:53:44 GMT (1530kb,D)

Link back to: arXiv, form interface, contact.