Two-shot learning of multiple strange attractors

2026/01/28 by Daniel Köglmayr, Miralem Spahic, Andrew Flynn +1 · 1 voice
Physics and Astronomy · #nlin.CD

paper · pdf · doi:10.1016/j.neunet.2026.109209

Abstract

The brain combines short- and long-term memory to process, store, and recall multiple different pieces of information. Inspired by this and recent results on multifunctional and parameter-aware learning, we extend a new machine learning technique that combines short- and long-term memory units, specifically, a system consisting of a next-generation reservoir computer (NGRC) and extremely randomized trees (ERT), to process, store, and recall multiple different strange attractors. We train the combined NGRC+ERT system using a two-shot learning approach which significantly improves performance by filtering out unnecessary features, thereby avoiding extensive hyperparameter optimization. We first show that an NGRC+ERT system achieves highly accurate reconstruction of the short- and long-term dynamics of both the Lorenz and Halvorsen chaotic attractors when using an exponential filtering scheme. We validate these finding by training the NGRC+ERT system to reconstruct different pairs of attractors and also a greater number of attractors. We focus on the task of training a single NGRC+ERT system to reconstruct 16 different attractors and show that sufficient index-based separation in feature space suppresses unwanted switching dynamics, thus stabilizing long-term memory recall. Finally, we identify that defects in short-term memory processing can provoke failure modes in long-term memory recall resulting in confabulation.

Citations

• Arousal as a universal embedding for spatiotemporal brain dynamics
• Next-generation reservoir computing for dynamical inference
• Multifunctional physical reservoir computing in soft tensegrity robots
• Confabulation dynamics in a reservoir computer: Filling in the gaps with untrained attractors
• Attractor-merging Crises and Intermittency in Reservoir Computing
• Controlling dynamical systems into unseen target states using machine learning
• Exploring the origins of switching dynamics in a multifunctional reservoir computer
• How more data can hurt: Instability and regularization in next-generation reservoir computing
• Tree-based Learning for High-Fidelity Prediction of Chaos
• Extrapolating tipping points and simulating non-stationary dynamics of complex systems using efficient machine learning
• Controlling dynamical systems to complex target states using machine learning: next-generation vs. classical reservoir computing
• Multifunctionality in a Connectome-Based Reservoir Computer
• Seeing double with a multifunctional reservoir computer
• Exploring the limits of multifunctionality across different reservoir computers
• Machine learning prediction of critical transition and system collapse
• A practical method for calculating largest Lyapunov exponents from small data sets
• Measuring the strangeness of strange attractors

Discussions

• Daniel K\"oglmayr, Miralem Spahic, Andrew Flynn, Christoph R\"ath: Two-shot learning of multiple strange attractors https://arxiv.org/abs/2601.21117 https://arxiv.org/pdf/2601.21117 https://arxiv.org/ [bsky, 1 points, 0 comments]