Learning continuous chaotic attractors with a reservoir computer

Abstract

Neural systems are well known for their ability to learn and store information as memories. Even more impressive is their ability to abstract these memories to create complex internal representations, enabling advanced functions such as the spatial manipulation of mental representations. While recurrent neural networks (RNNs) are capable of representing complex information, the exact mechanisms of how dynamical neural systems perform abstraction are still not well-understood, thereby hindering the development of more advanced functions. Here, we train a 1000-neuron RNN—a reservoir computer (RC)—to abstract a continuous dynamical attractor memory from isolated examples of dynamical attractor memories. Furthermore, we explain the abstraction mechanism with a new theory. By training the RC on isolated and shifted examples of either stable limit cycles or chaotic Lorenz attractors, the RC learns a continuum of attractors as quantified by an extra Lyapunov exponent equal to zero. We propose a theoretical mechanism of this abstraction by combining ideas from differentiable generalized synchronization and feedback dynamics. Our results quantify abstraction in simple neural systems, enabling us to design artificial RNNs for abstraction and leading us toward a neural basis of abstraction.

Document Details

Document Type
Pub Defense Publication
Publication Date
Jan 01, 2022
Source ID
10.1063/5.0075572

Entities

People

  • Danielle Bassett
  • Jason Z Kim
  • Lindsay M. Smith
  • Zhixin Lu

Organizations

  • Army Research Office
  • National Science Foundation
  • Office of Extramural Research
  • Paul G. Allen Family Foundation
  • Santa Fe Institute
  • University of Pennsylvania

Tags

Readers

  • Mathematical Modeling and Probability Theory.
  • Systems Analysis and Design
  • Wave Propagation and Nonlinear Chaotic Dynamics.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks