Look and Feel What and How Recurrent Self-Organizing Maps Learn

Jérémy Fix (1, 2, 3) , Hervé Frezza-Buet (1, 2, 3)

1 LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
2 CentraleSupélec
3 BISCUIT - Bio-Inspired, Situated and Cellular Unconventional Information Technologies

This paper introduces representations and measurements for revealing the inner self-organization that occurs in a 1D recurrent self-organizing map. Experiments show the incredible richness and robustness of an extremely simple architecture when it extracts hidden states of the HMM that feeds it with ambiguous and noisy inputs.

Mots clés: Recurrent self-organizing map; Sequence processing; Hidden markov models

Domaines: Réseau de neurones [cs.NE]

Jérémy Fix, Hervé Frezza-Buet. Look and Feel What and How Recurrent Self-Organizing Maps Learn. Advances in Self-Organizing Maps, Learning Vector Quantization, Clustering and Data Visualization, WSOM 19, 976, pp.3-12, 2020, Advances in Intelligent Systems and Computing, 978-3-030-19641-7. ⟨10.1007/978-3-030-19642-4_1⟩. ⟨hal-02120117⟩