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

Properties of incremental projection learning.

M Sugiyama1, H Ogawa

  • 1Department of Computer Science, Graduate School of Information Science and Engineering, Tokyo Institute of Technology, Japan. sugi@og.cs.titech.ac.jp

Neural Networks : the Official Journal of the International Neural Network Society
|February 24, 2001
PubMed
Summary

This study analyzes incremental projection learning, revealing that seemingly redundant data can improve future generalization. An improved redundancy criterion and a simpler, memory-efficient representation are proposed.

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

Atopic dermatitis: studies of skin permeability and effectiveness of topical PUVA treatment.

Pediatric dermatology·1992
Same author

Stimulation of keratinocyte migration by growth factors.

The Journal of dermatology·1992
Same author

A wound healing model using healing-impaired diabetic mice.

The Journal of dermatology·1992
Same author

The response to drug therapy in unstable angina on the basis of coronary angiography findings.

Japanese circulation journal·1992
Same author

[The pathogenic role of Chlamydia trachomatis in otitis media with effusion].

Nihon Jibiinkoka Gakkai kaiho·1992
Same author

[Prospective randomized trials comparing hyperfractionated radiotherapy with conventional radiotherapy in stage III non-small cell lung cancer].

Nihon Igaku Hoshasen Gakkai zasshi. Nippon acta radiologica·1992

Area of Science:

  • Machine Learning
  • Artificial Intelligence

Background:

  • Incremental learning methods often discard data considered redundant.
  • The properties of incremental projection learning require further investigation.

Purpose of the Study:

  • To analyze the properties of incremental projection learning.
  • To identify the potential effectiveness of seemingly redundant training data.
  • To develop an improved criterion for data redundancy and a more efficient representation.

Main Methods:

  • Analysis of training data effectiveness in incremental projection learning.
  • Derivation of a new criterion for identifying redundant training data.
  • Investigation of the relationship between prior and posterior learning results.
  • Classification of effective training data based on generalization improvement.

Related Experiment Videos

Main Results:

  • Seemingly redundant training data can significantly contribute to future generalization.
  • An improved criterion for data redundancy is derived.
  • Effective training data is categorized based on its impact on generalization.
  • A simplified representation of incremental projection learning is presented, requiring fixed memory.

Conclusions:

  • Incremental projection learning offers the same generalization capability as batch methods.
  • The proposed methods enhance understanding and efficiency in incremental learning.
  • The fixed-size memory representation makes the method scalable and practical.