ClassroomThe classroom
Back to the lesson libraryMECHANISM · 14 MIN
M37.6 CONNECT THE MECHANISM

Use fixed recurrent dynamics and train a readout

Build a recurrent network with random weights, never train them, and fit only the final layer. It forecasts a chaotic signal, and the whole training step is one line of linear algebra.

LESSON OVERVIEW14 min lesson

Lesson overview

Build a recurrent network with random weights, never train them, and fit only the final layer. It forecasts a chaotic signal, and the whole training step is one line of linear algebra.

What you’ll explore

  • Reservoir computing drives a dynamical system with inputs and learns an output mapping; memory, stability, observability, and readout training determine useful sequence processing.
GO TO THE SOURCE

Original explanations, connected to the research.

A Practical Guide to Applying Echo State Networks (Lukoševičius, 2012)Harnessing Nonlinearity: Predicting Chaotic Systems and Saving Energy in Wireless Communication (Jaeger & Haas, 2004)Neuronal Dynamics
Suggest a correction

A precise note can make an explanation better.

Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.

The file includes this note, the scene title, and lesson metadata. Your saved progress and quiz responses are excluded. Download before leaving or reloading to keep your note.