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

Incremental online learning in high dimensions.

Sethu Vijayakumar1, Aaron D'Souza, Stefan Schaal

  • 1School of Informatics, University of Edinburgh, Edinburgh EH9 3JZ, UK. sethu.vijayakumar@ed.ac.uk

Neural Computation
|October 11, 2005
PubMed
Summary
This summary is machine-generated.

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

RoboBallet: Planning for multirobot reaching with graph neural networks and reinforcement learning.

Science robotics·2025
Same author

Adaptive assistive robotics: a framework for triadic collaboration between humans and robots.

Royal Society open science·2023
Same author

A Haptic Sleeve as a Method of Mechanotactile Feedback Restoration for Myoelectric Hand Prosthesis Users.

Frontiers in rehabilitation sciences·2022
Same author

Action planning and control under uncertainty emerge through a desirability-driven competition between parallel encoding motor plans.

PLoS computational biology·2021
Same author

HapFIC: An Adaptive Force/Position Controller for Safe Environment Interaction in Articulated Systems.

IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society·2021
Same author

Effectively Quantifying the Performance of Lower-Limb Exoskeletons Over a Range of Walking Conditions.

Frontiers in robotics and AI·2021
Same journal

A Model-Free Reinforcement Learning Implementation of Decision Making Under Uncertainty by Sequential Sampling.

Neural computation·2026
Same journal

DROP: Distributional and Regular Optimism and Pessimism for Reinforcement Learning.

Neural computation·2026
Same journal

Hierarchical Active Inference Using Successor Representations.

Neural computation·2026
Same journal

W-Kernel and Its Principal Space for Frequentist Evaluation of Bayesian Estimators.

Neural computation·2026
Same journal

A Hidden Markov Model-Inspired Sequence Classification Method for Hyperdimensional Computing.

Neural computation·2026
Same journal

Sparse Graphical Modeling for Electrophysiological Phase-Based Connectivity Using Circular Statistics.

Neural computation·2026
See all related articles

Locally Weighted Projection Regression (LWPR) is a novel algorithm for incremental nonlinear function approximation. It efficiently handles high-dimensional data by using local linear models and partial least squares regression, offering robust and rapid learning.

Area of Science:

  • Machine Learning
  • Nonlinear Function Approximation
  • High-Dimensional Data Analysis

Background:

  • High-dimensional data presents challenges for traditional regression models.
  • Incremental learning methods are crucial for adaptive systems.
  • Existing methods struggle with redundant and irrelevant input dimensions.

Purpose of the Study:

  • Introduce and derive the Locally Weighted Projection Regression (LWPR) algorithm.
  • Demonstrate LWPR's effectiveness in high-dimensional, incremental nonlinear function approximation.
  • Provide a computationally efficient and numerically robust learning method.

Main Methods:

  • Nonparametric regression utilizing locally linear models.
  • Integration of partial least squares regression for dimensionality reduction.

Related Experiment Videos

  • Stochastic leave-one-out cross-validation for efficient learning.
  • Second-order learning methods for rapid incremental training.

    • Main Results:

    • LWPR demonstrates rapid learning with incremental, second-order methods.
    • The algorithm employs local information for weighting kernels, minimizing negative interference.
    • LWPR exhibits linear computational complexity concerning the number of inputs.
    • Empirical evaluations show successful application to datasets up to 90 dimensions.

    Conclusions:

    • LWPR is a pioneering incremental, spatially localized learning method for high-dimensional spaces.
    • The algorithm offers computational efficiency and numerical robustness.
    • LWPR successfully handles redundant and irrelevant input dimensions in nonlinear function approximation.