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

Parameter learning from stochastic teachers and stochastic compulsive liars.

B John Oommen1, Govindachari Raghunath, Benjamin Kuipers

  • 1School of Computer Science, Carleton University, Ottawa, ON, Canada. oommen@scs.carleton.ca

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|August 15, 2006
PubMed
Summary

This study introduces a new learning strategy for robots and algorithms to find unknown parameters in uncertain environments. The CPL-AdS method efficiently narrows down the search space, even when the feedback source is unreliable or deceptive.

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

The Hierarchical Discrete Pursuit Learning Automaton: A Novel Scheme With Fast Convergence and Epsilon-Optimality.

IEEE transactions on neural networks and learning systems·2023
Same author

Trust and Cooperation.

Frontiers in robotics and AI·2022
Same author

Solving Two-Person Zero-Sum Stochastic Games With Incomplete Information Using Learning Automata With Artificial Barriers.

IEEE transactions on neural networks and learning systems·2021
Same author

Achieving Fair Load Balancing by Invoking a Learning Automata-Based Two-Time-Scale Separation Paradigm.

IEEE transactions on neural networks and learning systems·2020
Same author

The Hierarchical Continuous Pursuit Learning Automation: A Novel Scheme for Environments With Large Numbers of Actions.

IEEE transactions on neural networks and learning systems·2019
Same author

A Conclusive Analysis of the Finite-Time Behavior of the Discretized Pursuit Learning Automaton.

IEEE transactions on neural networks and learning systems·2019

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Control Theory

Background:

  • Traditional learning automata (LA) models assume optimal action selection from an environment.
  • This paper addresses learning a parameter within an interval, interacting with a stochastic environment.
  • The environment may provide feedback with a probability p, being informative (p > 0.5) or deceptive (p < 0.5).

Purpose of the Study:

  • To develop a novel learning strategy for a mechanism to learn an unknown parameter in a closed interval.
  • To enable learning from both stochastic teachers and stochastic compulsive liars.
  • To ensure the strategy works effectively even when the mechanism is unaware of the environment's feedback accuracy.

Main Methods:

  • The proposed CPL-AdS algorithm partitions the search interval into subintervals.

Related Experiment Videos

  • It utilizes fast-converging E-optimal LRI LA to evaluate the unknown point's location.
  • The search space is pruned iteratively by eliminating at least one partition.

    • Main Results:

    • The CPL-AdS algorithm provably converges to the unknown point with desired accuracy.
    • It achieves high probability of success, approaching unity.
    • Experimental results confirm fast and accurate convergence across various feedback accuracy (p) values.

    Conclusions:

    • The CPL-AdS strategy offers a robust solution for parameter learning in uncertain and potentially deceptive environments.
    • This is the first reported method applicable to learning from stochastic compulsive liars.
    • The algorithm's adaptability and proven convergence suggest numerous potential applications.