RNNs perform task computations by dynamically warping neural representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Pellegrino, Arthur, Chadwick, Angus
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917256483569664
author Pellegrino, Arthur
Chadwick, Angus
author_facet Pellegrino, Arthur
Chadwick, Angus
contents Analysing how neural networks represent data features in their activations can help interpret how they perform tasks. Hence, a long line of work has focused on mathematically characterising the geometry of such "neural representations." In parallel, machine learning has seen a surge of interest in understanding how dynamical systems perform computations on time-varying input data. Yet, the link between computation-through-dynamics and representational geometry remains poorly understood. Here, we hypothesise that recurrent neural networks (RNNs) perform computations by dynamically warping their representations of task variables. To test this hypothesis, we develop a Riemannian geometric framework that enables the derivation of the manifold topology and geometry of a dynamical system from the manifold of its inputs. By characterising the time-varying geometry of RNNs, we show that dynamic warping is a fundamental feature of their computations.
format Preprint
id arxiv_https___arxiv_org_abs_2512_04310
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RNNs perform task computations by dynamically warping neural representations
Pellegrino, Arthur
Chadwick, Angus
Machine Learning
Differential Geometry
Dynamical Systems
Neurons and Cognition
Analysing how neural networks represent data features in their activations can help interpret how they perform tasks. Hence, a long line of work has focused on mathematically characterising the geometry of such "neural representations." In parallel, machine learning has seen a surge of interest in understanding how dynamical systems perform computations on time-varying input data. Yet, the link between computation-through-dynamics and representational geometry remains poorly understood. Here, we hypothesise that recurrent neural networks (RNNs) perform computations by dynamically warping their representations of task variables. To test this hypothesis, we develop a Riemannian geometric framework that enables the derivation of the manifold topology and geometry of a dynamical system from the manifold of its inputs. By characterising the time-varying geometry of RNNs, we show that dynamic warping is a fundamental feature of their computations.
title RNNs perform task computations by dynamically warping neural representations
topic Machine Learning
Differential Geometry
Dynamical Systems
Neurons and Cognition
url https://arxiv.org/abs/2512.04310