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

Reinforcement learning control

A G Barto1

  • 1Department of Computer Science, University of Massachusetts, Amherst 01003, USA.

Current Opinion in Neurobiology
|December 1, 1994
PubMed
Summary
This summary is machine-generated.

Reinforcement learning enables autonomous systems to learn from experience, bypassing the need for human teachers. This trial-and-error approach is crucial when training data is scarce or unavailable.

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

Distributed representation of limb motor programs in arrays of adjustable pattern generators.

Journal of cognitive neuroscience·2013
Same author

A cerebellar model of timing and prediction in the control of reaching.

Neural computation·1999
Same author

Prediction of complex two-dimensional trajectories by a cerebellar model of smooth pursuit eye movement.

Journal of neurophysiology·1997
Same author

Distributed motor commands in the limb premotor network.

Trends in neurosciences·1993
Same author

Linear systems analysis of the relationship between firing of deep cerebellar neurons and the classically conditioned nictitating membrane response in rabbits.

Biological cybernetics·1991
Same author

Simulation of the classically conditioned nictitating membrane response by a neuron-like adaptive element: response topography, neuronal firing, and interstimulus intervals.

Behavioural brain research·1986

Area of Science:

  • Artificial Intelligence
  • Computational Neuroscience
  • Machine Learning

Background:

  • Most artificial learning systems rely on 'teacher' guidance for training stimuli.
  • This supervised learning approach is inadequate when training data is costly or impossible to obtain.
  • Reinforcement learning (RL) offers an alternative by enabling learning from experience.

Purpose of the Study:

  • To highlight the growing importance of reinforcement learning in AI and computational neuroscience.
  • To explain the principles of RL as a method for autonomous learning.
  • To discuss the advantages of RL over traditional supervised learning methods.

Main Methods:

  • Exploration of trial-and-error learning paradigms.
  • Application of RL principles to autonomous systems.

Related Experiment Videos

  • Alignment of RL with biological learning principles.

    • Main Results:

    • Recent advancements have significantly improved the efficiency of reinforcement learning algorithms.
    • RL allows autonomous systems to learn effectively without direct instruction.
    • The approach aligns with biological learning mechanisms, making it relevant for neuroscience.

    Conclusions:

    • Reinforcement learning is a vital tool for developing autonomous systems capable of learning from their environment.
    • Its efficiency and biological plausibility are driving increased interest in computer science, engineering, and neuroscience.
    • RL represents a significant step towards more adaptable and independent artificial intelligence.