Related Experiment Video
Updated: Aug 6, 2026

Morphology-Based Distinction Between Healthy and Pathological Cells Utilizing Fourier Transforms and Self-Organizing Maps
Published on: October 28, 2018
Dynamics and topographic organization of recursive self-organizing maps
Peter Tino1, Igor Farkas, Jort van Mourik
1School of Computer Science, University of Birmingham, Birmingham B15 2TT, UK. P.Tino@cs.bham.ac.uk
Researchers analyzed the recursive self-organizing map (RecSOM) for sequential data processing. They found that trainable feedback connections enhance memory depth and topography preservation in Markovian topographic maps.
Area of Science:
- Neurocomputation
- Dynamical Systems
- Machine Learning
Background:
- Extending topographic maps to sequential data is an active research area.
- Current models lack clear understanding of representational capabilities and internal representations.
- General consensus on optimal sequence processing with topographic maps is missing.
Purpose of the Study:
- To rigorously analyze the recursive self-organizing map (RecSOM) for sequential data.
- To investigate the conditions for generating Markovian organizations of receptive fields.
- To explore methods for improving memory depth and topography preservation in sequential data mapping.
Main Methods:
- Analysis of RecSOM as a nonautonomous dynamical system with fixed input maps.
- Derivation of parameter bounds for guaranteeing contractiveness of fixed-input maps.
- Comparison of fixed dynamic modules versus trainable feedback connections in RecSOM.
Main Results:
- Contractive fixed-input maps are likely to produce Markovian organizations.
- Bounds for parameter beta ensuring contractiveness were derived.
- Trainable feedback connections yield superior memory depth and topography preservation compared to fixed modules.
Conclusions:
- RecSOM provides a framework for Markovian topographic maps of sequential data.
- Fixed dynamic modules can create Markovian maps, but trainable feedback offers enhanced performance.
- Non-Markovian organizations are important for advanced sequential data topographic mapping.
Related Concept Videos
Plotting of Topographic Maps
Level Curves and Contour Maps
Methods of Obtaining Topography
Topographic Surveying and Contours
Woodward–Hoffmann Selection Rules and Microscopic Reversibility
Manipulation and Analysis
