NP-SOM: network programmable self-organizing maps - Archive ouverte HAL Access content directly
Conference Papers Year :

NP-SOM: network programmable self-organizing maps

(1, 2) , (3) , (1) , (1)
1
2
3

Abstract

Self-organizing maps (SOM) are a well-known and biologically plausible model of input-driven self-organization that has shown to be effective in a wide range of applications. We want to use SOMs to control the processing cores of a massively parallel digital reconfigurable hardware, taking into account the communication constraints of its underlying network-on-chip (NoC) thanks to bio-inspired principles of structural plasticity. Although the SOM accounts for synaptic plasticity, it doesn't address structural plasticity. Therefore we have developed a model, namely the NP-SOM (network programmable self-organizing map), able to define SOMs with different underlying topologies as the result of a specific configuration of the associated NoC. To gain insights on a future introduction of advanced structural plasticity rules that will induce dynamic topological modifications, we investigate and quantify the effects of different hardware-compatible topologies on the SOM performance. To perform our tests we consider a lossy image compression as an illustrative application.
Fichier principal
Vignette du fichier
NP_SOM__network_programmable_self_organizing_maps-2.pdf (772.01 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02058458 , version 1 (06-03-2019)

Identifiers

Cite

Yann Bernard, Emeline Buoy, Adrien Fois, Bernard Girau. NP-SOM: network programmable self-organizing maps. 2018 IEEE 30th International Conference on Tools with Artificial Intelligence (ICTAI), Nov 2018, Volos, Greece. pp.908-915, ⟨10.1109/ICTAI.2018.00141⟩. ⟨hal-02058458⟩
62 View
128 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More