Efficient Implementations of Echo State Network Cross-Validation

Lukoševičius, Mantas; Uselis, Arnas

doi:10.1007/s12559-021-09849-2

Cited by 17 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following a convention adopted in the literature on ESN and relevant ML algorithms [26], as the first test time series, we employ a chaotic Mackey-Glass time series (MGTS) [77]. Other chaotic time series of artificial [16,35,[78][79][80][81] and natural origin [16,82] have also been employed to test ESN similarly to what is presented below.…”

Section: Examples Of Operation Of Esnmentioning

confidence: 99%

See 1 more Smart Citation

Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond

Maksymov

2023

Energies

View full text Add to dashboard Cite

More than 3.5 billion people live in rural areas, where water and water energy resources play an important role in ensuring sustainable and productive rural economies. This article reviews and critically analyses the recent advances in the field of analogue and reservoir computing that have been driven by the unique physical properties and energy of water waves. It also demonstrates that analogue and physical reservoir computing, taken as an independent research field, holds the potential to bring artificial intelligence closer to people living outside large cities, thus enabling them to enjoy the benefits of novel technologies that are already in place in large cities but are not readily available or suitable for regional communities. In particular, although the physical reservoir computing systems discussed in the main text are universal in terms of processing input data and making forecasts, they can be used to design and optimise power grid networks and forecast energy consumption, both at local and global scales. Thus, this review article will be of interest to a broad readership interested in novel concepts of artificial intelligence and machine learning and their innovative practical applications in diverse areas of science and technology.

show abstract

Section: Examples Of Operation Of Esnmentioning

confidence: 99%

“…The resulting interference patterns were projected and recorded at a resolution of 320 × 240 PX at 5 frames per second rate (Figure 6b). Each frame was digitally processed to remove noise and to obtain 50 virtual neurons (see [82,84] for a relevant discussion).…”

Section: Pioneering Workmentioning

confidence: 99%

Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond

Maksymov

2023

Energies

View full text Add to dashboard Cite

show abstract

“…This class of models is directly derived from the Boussinesq equations of two-dimensional thermal convection between two impermeable parallel plates, heated uniformly from below and cooled from above with free-slip boundary conditions for the velocity field [44,45]. Here, we explore QRCMs in two different modes of operation [46]: Secondly, we directly compare the results of the QRCM to its classical counterpart for the same flow. We identify hyperparameters in both approaches that can be related to each other.…”

Section: Introductionmentioning

confidence: 99%

Hybrid quantum-classical reservoir computing of thermal convection flow

Pfeffer¹,

Heyder²,

Schumacher³

2022

Preprint

View full text Add to dashboard Cite

We simulate the nonlinear chaotic dynamics of Lorenz-type models for a classical two-dimensional thermal convection flow with 3 and 8 degrees of freedom by a hybrid quantum-classical reservoir computing model. The high-dimensional quantum reservoir dynamics is established by universal quantum gates that rotate and entangle the individual qubits of the tensor product quantum state. A comparison of the quantum reservoir computing model with its classical counterpart shows that the same prediction and reconstruction capabilities of classical reservoirs with thousands of perceptrons can be obtained by a few strongly entangled qubits. We demonstrate that the mean squared error between model output and ground truth in the test phase of the quantum reservoir computing algorithm increases when the reservoir is decomposed into separable subsets of qubits. Furthermore, the quantum reservoir computing model is implemented on a real noisy IBM quantum computer for up to 7 qubits. Our work thus opens the door to model the dynamics of classical complex systems in a high-dimensional phase space effectively with an algorithm that requires a small number of qubits.

show abstract

“…Recording the intensity of this pixel as a function of time and processing this dependence using a computer, we obtain a reservoir with just one physical node. While the feasibility of single-node reservoirs has been demonstrated [24], we follow another standard approach that is well-known in the fields of both algorithmic [25] and physical RC [26,27], where at the post-processing stage we produce virtual neurons by sampling the experimentally measured response of SL waves. We empirically established that a practicable reservoir should contain 20-40 virtual neurons and that a further increase in the number of neurons results in a gradual improvement of the reservoir performance.…”

mentioning

confidence: 99%

Reservoir computing based on solitary-like waves dynamics of liquid film flows: A proof of concept

Maksymov

Pototsky

2023

EPL

View full text Add to dashboard Cite

Several theoretical works have shown that solitons---waves that self-maintain constant shape and velocity as they propagate---can be used as a physical computational reservoir, a concept where machine learning algorithms designed for digital computers are replaced by analog physical systems that exhibit nonlinear dynamical behaviour. Here we propose and experimentally validate a novel reservoir computing (RC) system that for the first time employs solitary-like (SL) waves propagating on the surface of a liquid film flowing over an inclined surface. We demonstrate the ability of the SL wave RC system (SLRC) to forecast chaotic time series, also producing experimental evidence of the possibility to combine RC with nonlinear vector autoregression techniques not only in a computer program but also in a physical systems.

show abstract

Efficient Implementations of Echo State Network Cross-Validation

Cited by 17 publications

References 18 publications

Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond

Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond

Hybrid quantum-classical reservoir computing of thermal convection flow

Reservoir computing based on solitary-like waves dynamics of liquid film flows: A proof of concept

Contact Info

Product

Resources

About