BROWSE

ITEM VIEW & DOWNLOAD

Enhanced reservoir computing with concatenation

Cited 0 times inthomson ciCited 0 times inthomson ci
Title
Enhanced reservoir computing with concatenation
Author
Park, Eonyoung
Advisor
Jang, Bongsoo
Issue Date
2020-08
Publisher
Graduate School of UNIST
Abstract
Recently, Reservoir Computing(RC) is spotlighted because of structural simplicity. RC is an architecture based on a Recurrent Neural Network(RNN) and the recurrent connections in RNN is called "reservoir" in RC. In this paper, we introduce the echo state network(ESN), which is one of the representative models of RC, and advanced versions of ESN. We introduce the dynamic based RC with the Kuramoto model(RCK), which is a mathematical model used to describe synchronization. There are many modified versions of Kuramoto model, we use the explosive Kuramoto model in this paper. We test RCK using the tasks in the advanced version of ESN articles, and compare RCK to a simplified ESN. We propose the concatenated RC. We apply the concatenation to RC, and compare the effect of concatenation using the capacit
Description
Department of Mathematical Sciences
URI
Go to Link
Appears in Collections:
MTH_Theses_Master
Files in This Item:
Enhanced reservoir computing with concatenation.pdf Download

find_unist can give you direct access to the published full text of this article. (UNISTARs only)

Show full item record

qrcode

  • mendeley

    citeulike

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

MENU