State space models, emergence, and ergodicity: How many parameters are needed for stable predictions?

TL;DR

This paper demonstrates that in simple linear dynamical systems, a critical number of parameters is necessary for stable, long-range predictions, revealing an emergence-like phase transition in model complexity.

Contribution

It shows that a phase transition exists in learning linear dynamical systems, linking parameter count to the ability to achieve bounded error in long sequences.

Findings

Non-ergodic systems require a threshold number of parameters for stable predictions.

Tasks with long-range correlations demand a critical number of parameters, akin to emergence.

Linear filters cannot effectively learn hidden state models unless their length exceeds a certain threshold.

Abstract

How many parameters are required for a model to execute a given task? It has been argued that large language models, pre-trained via self-supervised learning, exhibit emergent capabilities such as multi-step reasoning as their number of parameters reach a critical scale. In the present work, we explore whether this phenomenon can analogously be replicated in a simple theoretical model. We show that the problem of learning linear dynamical systems -- a simple instance of self-supervised learning -- exhibits a corresponding phase transition. Namely, for every non-ergodic linear system there exists a critical threshold such that a learner using fewer parameters than said threshold cannot achieve bounded error for large sequence lengths. Put differently, in our model we find that tasks exhibiting substantial long-range correlation require a certain critical number of parameters -- a phenomenon akin to emergence. We also investigate the role of the learner's parametrization and consider a simple version of a linear dynamical system with hidden state -- an imperfectly observed random walk in $\mathbb{R}$. For this situation, we show that there exists no learner using a linear filter which can succesfully learn the random walk unless the filter length exceeds a certain threshold depending on the effective memory length and horizon of the problem.