Reservoir computing

Machine learning, Unconventional computing, Unsupervised learning

Reservoir computing is a term used to describe a class of machine learning algorithms that rely on transient dynamics of a dynamical system to implement and manipulate goal-related information.

The most famous example is echo-state networks, which uses random recurrent neural networks as reservoirs, but other dynamical systems can also be used.

Reservoir computing with cellular automata

Reservoir computing can use cellular automata as the reservoir. Some citations (Nichele, Molund 2017; Yilmaz 2014; Morán et al. 2018; Babson et al. 2019).

Echo-state networks

Reservoir computing for differential equation solving

(Mattheakis et al. 2021)

Reservoir computing in physical media

(Tanaka et al. 2019)


  1. . . "Deep Reservoir Computing Using Cellular Automata". Arxiv:1703.02806 [cs].
  2. . . "Reservoir Computing Using Cellular Automata". Arxiv:1410.0162 [cs].
  3. . . "Reservoir Computing Hardware with Cellular Automata". Arxiv:1806.04932 [nlin].
  4. . . "Reservoir Computing with Complex Cellular Automata". Complex Systems 28 (4):433–55. DOI.
  5. . . "Unsupervised Reservoir Computing for Solving Ordinary Differential Equations". Arxiv:2108.11417 [physics].
  6. . . "Recent Advances in Physical Reservoir Computing: A Review". Neural Networks 115 (July):100–123. DOI.
Last changed | authored by


← Back to Notes