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

Reducing topological defects in self-organizing maps using multiple scale neighborhood functions.

Kazushi Murakoshi1, Yuichi Sato

  • 1Department of Knowledge-based Information Engineering, Toyohashi University of Technology, 1-1 Hibarigaoka, Tenpaku-cho, Toyohashi 441-8580, Japan. mura@tutkie.tut.ac.jp

Bio Systems
|September 12, 2006
PubMed
Summary

We introduce a novel method using multi-scale neighborhood functions to reduce topological defects in self-organizing maps (SOMs). This approach improves solutions for the Traveling Salesman Problem (TSP), lowering tour length and solution kinks.

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

Association Between CKAP4 Expression and Poor Prognosis in Patients with Bladder Cancer Treated with Radical Cystectomy.

Cancers·2025
Same author

Predictors of prolonged severe urinary incontinence after robot-assisted radical prostatectomy: a prospective study.

Fukushima journal of medical science·2025
Same author

Effects of Sacubitril/Valsartan on Myocardial Tissue in Heart Failure With Left Ventricular Ejection Fraction Below 50.

Circulation journal : official journal of the Japanese Circulation Society·2025
Same author

De novo overactive bladder after robot-assisted radical prostatectomy is predicted by a high preoperative overactive bladder symptom score, preoperative dyslipidemia, and non-nerve-sparing surgery.

Fukushima journal of medical science·2025
Same author

Serum low density lipoprotein cholesterol / high density lipoprotein cholesterol ratio predicts lower urinary tract dysfunction related dyslipidemia.

Fukushima journal of medical science·2025
Same author

Early Rehabilitation Interventions by Physical Therapists for Severe COVID-19 Patients Were Associated With Decreased Incidence of Post-ICU Physical Impairment.

Annals of rehabilitation medicine·2025

Area of Science:

  • Artificial Intelligence
  • Computational Neuroscience
  • Machine Learning

Background:

  • Self-organizing maps (SOMs) are powerful unsupervised learning algorithms used for dimensionality reduction and data visualization.
  • Topological defects can arise in SOMs, potentially hindering the accurate representation of data topology.
  • Existing methods for defect reduction may have limitations in preserving global data structures.

Purpose of the Study:

  • To propose and evaluate a novel method for reducing topological defects in self-organizing maps.
  • To leverage principles from the human visual system's multi-scale processing for improved SOM performance.
  • To demonstrate the efficacy of the proposed method on a complex optimization problem.

Main Methods:

  • Development of a SOM algorithm incorporating multiple scale neighborhood functions, inspired by human visual processing.

Related Experiment Videos

  • Application of the proposed SOM method to the Traveling Salesman Problem (TSP).
  • Quantitative analysis of solution quality using tour length and the number of kinks as evaluation metrics.

    • Main Results:

    • The proposed method demonstrated a reduction in both the tour length and the number of kinks in TSP solutions compared to baseline methods.
    • The multi-scale neighborhood functions effectively mitigated topological defects within the self-organizing map.
    • Consistent improvements were observed across the evaluated metrics, indicating enhanced solution quality.

    Conclusions:

    • The proposed multi-scale neighborhood function approach is effective in reducing topological defects in SOMs.
    • This method offers a promising strategy for improving the performance of SOMs in complex problem-solving tasks.
    • The findings suggest a potential link between biological visual processing and artificial neural network optimization.