Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Grounding words in perception and action: computational insights.

Deb Roy1

  • 1The Media Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02142, USA. dkroy@media.mit.edu

Trends in Cognitive Sciences
|July 12, 2005
PubMed
Summary

Researchers are developing computational models that ground language in robotic perception and action. These systems bridge symbolic language with the physical world, explaining context-dependent word meanings and offering new cognitive modeling methods.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Language model estimates indicate increased outgroup animosity among a sample of politicians, journalists, and partisan users on Twitter and Reddit.

Scientific reports·2026
Same author

The speed of news in Twitter (X) versus radio.

Scientific reports·2024
Same author

Allograft nephrectomy versus nonallograft nephrectomy after failed renal transplantation: a systematic review by updated meta-analysis.

Transplant international : official journal of the European Society for Organ Transplantation·2021
Same author

Minimally invasive stabilization for thoracolumbar and lumbar fractures: a comparative study with short segment open Schanz screw constructs.

Journal of spine surgery (Hong Kong)·2019
Same author

Minimally invasive stabilization of the fractured ankylosed spine: a comparative case series study.

Journal of spine surgery (Hong Kong)·2018
Same author

The spread of true and false news online.

Science (New York, N.Y.)·2018

Area of Science:

  • Cognitive Science
  • Robotics
  • Computational Linguistics
  • Artificial Intelligence

Background:

  • Language is fundamental for communication about objects, properties, relations, and actions.
  • Traditional symbolic models struggle to explain context-dependent word meaning shifts.
  • Emerging research focuses on grounding language in machine perception and action.

Purpose of the Study:

  • To introduce a new computational model that bridges symbolic language with real-world referents.
  • To explain context-dependent word meaning shifts not easily addressed by purely symbolic approaches.
  • To explore the potential of grounded systems for cognitive modeling and situated learning research.

Main Methods:

  • Development of robotic and simulated systems capable of grounding language in perception and action.

Related Experiment Videos

  • Creation of novel computational models integrating symbolic representations with physical world interactions.
  • Utilizing first-person-perspective sensory data processing within grounded systems.

    • Main Results:

    • A new class of computational models has emerged, connecting language to physical reality.
    • These models successfully explain context-dependent variations in word meaning.
    • Grounded systems demonstrate the ability to process first-person sensory data.

    Conclusions:

    • Grounded language models offer a more comprehensive explanation for word meaning than purely symbolic systems.
    • These systems provide a novel methodology for testing cognitive hypotheses related to situated communication and learning.
    • The ability to process first-person sensory data opens new avenues for cognitive modeling research.