Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Quad-Q-learning.

C Clausen¹, H Wechsler

¹Department of Computer Science, George Mason University, Fairfax, VA 22030, USA.

IEEE Transactions on Neural Networks

|February 6, 2008

Summary

This summary is machine-generated.

Quad-Q-learning is a novel unsupervised learning algorithm for autonomous agents, extending Q-learning by enabling actions to branch into four new states. This hierarchical approach enhances learning efficiency for complex problems, including image compression.

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

Child and adolescent psychiatry training and mental health care in Southeast Europe.

European child & adolescent psychiatry·2019

Same author

SU-E-T-598: Clinical Experience of Configuration, Commission and Implementation for SmartArc with MOSAIQ R&V System.

Medical physics·2017

Same author

The uses and misuses of data in assessing mental health needs.

Community mental health journal·2013

Same author

A random walk procedure for texture discrimination.

IEEE transactions on pattern analysis and machine intelligence·2011

Same author

On the difficulties involved in the segmentation of pictures.

IEEE transactions on pattern analysis and machine intelligence·2011

Same author

A theory for invariant object recognition in the frontoparallel plane.

IEEE transactions on pattern analysis and machine intelligence·2011

Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013

Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011

Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011

Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011

Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011

Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011

See all related articles

Area of Science:

Artificial Intelligence
Machine Learning
Reinforcement Learning

Background:

Q-learning is a reinforcement learning algorithm that learns optimal policies by maximizing rewards.
Traditional Q-learning environments transition to a single new state after an action.
Complex problems often require more sophisticated state transition mechanisms for efficient learning.

Purpose of the Study:

To introduce quad-Q-learning, a new learning algorithm inspired by Q-learning.
To enable autonomous agents to learn optimally in environments with complex state transitions.
To develop a hierarchical learning approach for tackling large, intractable problem domains.

Main Methods:

Quad-Q-learning allows an action from a state to result in either an immediate reward or four new, independent states.

Related Experiment Videos

The algorithm operates in a hierarchical state space, processing lower-level states before higher levels.

Two versions are presented: discrete state and mixed discrete/continuous state quad-Q-learning, utilizing functional approximation for continuous states.

Main Results:

Quad-Q-learning improves learning efficiency at higher levels of the state hierarchy.
The algorithm can be generalized to n-Q-learning for partitioning large problems into independently solvable subproblems.
Application to image compression demonstrates practical utility, especially with continuous state learning.

Conclusions:

Quad-Q-learning offers a powerful framework for unsupervised learning in hierarchical environments.
The "divide and conquer" approach inherent in quad-Q-learning is crucial for scaling to complex, real-world problems.
Continuous state learning is essential for practical applications of quad-Q-learning, overcoming limitations of discrete state versions.