Existence of reservoir with finite-dimensional output for universal reservoir computing
© 2024. The Author(s)..
In this paper, we prove the existence of a reservoir that has a finite-dimensional output and makes the reservoir computing model universal. Reservoir computing is a method for dynamical system approximation that trains the static part of a model but fixes the dynamical part called the reservoir. Hence, reservoir computing has the advantage of training models with a low computational cost. Moreover, fixed reservoirs can be implemented as physical systems. Such reservoirs have attracted attention in terms of computation speed and energy consumption. The universality of a reservoir computing model is its ability to approximate an arbitrary system with arbitrary accuracy. Two sufficient reservoir conditions to make the model universal have been proposed. The first is the combination of fading memory and the separation property. The second is the neighborhood separation property, which we proposed recently. To date, it has been unknown whether a reservoir with a finite-dimensional output can satisfy these conditions. In this study, we prove that no reservoir with a finite-dimensional output satisfies the former condition. By contrast, we propose a single output reservoir that satisfies the latter condition. This implies that, for any dimension, a reservoir making the model universal exists with the output of that specified dimension. These results clarify the practical importance of our proposed conditions.
| Medienart: |
E-Artikel |
|---|
| Erscheinungsjahr: |
2024 |
|---|---|
| Erschienen: |
2024 |
| Enthalten in: |
Zur Gesamtaufnahme - volume:14 |
|---|---|
| Enthalten in: |
Scientific reports - 14(2024), 1 vom: 11. Apr., Seite 8448 |
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Sugiura, Shuhei [VerfasserIn] |
|---|
| Links: |
|---|
| Themen: |
Journal Article |
|---|
| Anmerkungen: |
Date Revised 25.04.2024 published: Electronic Citation Status PubMed-not-MEDLINE |
|---|
| doi: |
10.1038/s41598-024-56742-7 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
NLM370894812 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | NLM370894812 | ||
| 003 | DE-627 | ||
| 005 | 20240425233748.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 240411s2024 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1038/s41598-024-56742-7 |2 doi | |
| 028 | 5 | 2 | |a pubmed24n1386.xml |
| 035 | |a (DE-627)NLM370894812 | ||
| 035 | |a (NLM)38600157 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Sugiura, Shuhei |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Existence of reservoir with finite-dimensional output for universal reservoir computing |
| 264 | 1 | |c 2024 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a ƒaComputermedien |b c |2 rdamedia | ||
| 338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Date Revised 25.04.2024 | ||
| 500 | |a published: Electronic | ||
| 500 | |a Citation Status PubMed-not-MEDLINE | ||
| 520 | |a © 2024. The Author(s). | ||
| 520 | |a In this paper, we prove the existence of a reservoir that has a finite-dimensional output and makes the reservoir computing model universal. Reservoir computing is a method for dynamical system approximation that trains the static part of a model but fixes the dynamical part called the reservoir. Hence, reservoir computing has the advantage of training models with a low computational cost. Moreover, fixed reservoirs can be implemented as physical systems. Such reservoirs have attracted attention in terms of computation speed and energy consumption. The universality of a reservoir computing model is its ability to approximate an arbitrary system with arbitrary accuracy. Two sufficient reservoir conditions to make the model universal have been proposed. The first is the combination of fading memory and the separation property. The second is the neighborhood separation property, which we proposed recently. To date, it has been unknown whether a reservoir with a finite-dimensional output can satisfy these conditions. In this study, we prove that no reservoir with a finite-dimensional output satisfies the former condition. By contrast, we propose a single output reservoir that satisfies the latter condition. This implies that, for any dimension, a reservoir making the model universal exists with the output of that specified dimension. These results clarify the practical importance of our proposed conditions | ||
| 650 | 4 | |a Journal Article | |
| 650 | 4 | |a Machine learning | |
| 650 | 4 | |a Neural network | |
| 650 | 4 | |a Nonlinear dynamical system | |
| 650 | 4 | |a Reservoir computing | |
| 700 | 1 | |a Ariizumi, Ryo |e verfasserin |4 aut | |
| 700 | 1 | |a Asai, Toru |e verfasserin |4 aut | |
| 700 | 1 | |a Azuma, Shun-Ichi |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Scientific reports |d 2011 |g 14(2024), 1 vom: 11. Apr., Seite 8448 |w (DE-627)NLM215703936 |x 2045-2322 |7 nnns |
| 773 | 1 | 8 | |g volume:14 |g year:2024 |g number:1 |g day:11 |g month:04 |g pages:8448 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1038/s41598-024-56742-7 |3 Volltext |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a GBV_NLM | ||
| 951 | |a AR | ||
| 952 | |d 14 |j 2024 |e 1 |b 11 |c 04 |h 8448 | ||