Weight-Space Linear Recurrent Neural Networks

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Nzoyem, Roussel Desmond, Keshtmand, Nawid, Fernandez, Enrique Crespo, Tsayem, Idriss, Santos-Rodriguez, Raul, Barton, David A. W., Deakin, Tom
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910038394667008
author Nzoyem, Roussel Desmond
Keshtmand, Nawid
Fernandez, Enrique Crespo
Tsayem, Idriss
Santos-Rodriguez, Raul
Barton, David A. W.
Deakin, Tom
author_facet Nzoyem, Roussel Desmond
Keshtmand, Nawid
Fernandez, Enrique Crespo
Tsayem, Idriss
Santos-Rodriguez, Raul
Barton, David A. W.
Deakin, Tom
contents We introduce WARP (Weight-space Adaptive Recurrent Prediction), a simple yet powerful model that unifies weight-space learning with linear recurrence to redefine sequence modeling. Unlike conventional recurrent neural networks (RNNs) which collapse temporal dynamics into fixed-dimensional hidden states, WARP explicitly parametrizes its hidden state as the weights and biases of a distinct auxiliary neural network, and uses input differences to drive its recurrence. This brain-inspired formulation enables efficient gradient-free adaptation of the auxiliary network at test-time, in-context learning abilities, and seamless integration of domain-specific physical priors. Empirical validation shows that WARP matches or surpasses state-of-the-art baselines on diverse classification tasks, featuring in the top three in 4 out of 6 real-world challenging datasets. Furthermore, extensive experiments across sequential image completion, multivariate time series forecasting, and dynamical system reconstruction demonstrate its expressiveness and generalisation capabilities. Remarkably, a physics-informed variant of our model outperforms the next best model by more than 10x. Ablation studies confirm the architectural necessity of key components, solidifying weight-space linear RNNs as a transformative paradigm for adaptive machine intelligence.
format Preprint
id arxiv_https___arxiv_org_abs_2506_01153
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Weight-Space Linear Recurrent Neural Networks
Nzoyem, Roussel Desmond
Keshtmand, Nawid
Fernandez, Enrique Crespo
Tsayem, Idriss
Santos-Rodriguez, Raul
Barton, David A. W.
Deakin, Tom
Machine Learning
We introduce WARP (Weight-space Adaptive Recurrent Prediction), a simple yet powerful model that unifies weight-space learning with linear recurrence to redefine sequence modeling. Unlike conventional recurrent neural networks (RNNs) which collapse temporal dynamics into fixed-dimensional hidden states, WARP explicitly parametrizes its hidden state as the weights and biases of a distinct auxiliary neural network, and uses input differences to drive its recurrence. This brain-inspired formulation enables efficient gradient-free adaptation of the auxiliary network at test-time, in-context learning abilities, and seamless integration of domain-specific physical priors. Empirical validation shows that WARP matches or surpasses state-of-the-art baselines on diverse classification tasks, featuring in the top three in 4 out of 6 real-world challenging datasets. Furthermore, extensive experiments across sequential image completion, multivariate time series forecasting, and dynamical system reconstruction demonstrate its expressiveness and generalisation capabilities. Remarkably, a physics-informed variant of our model outperforms the next best model by more than 10x. Ablation studies confirm the architectural necessity of key components, solidifying weight-space linear RNNs as a transformative paradigm for adaptive machine intelligence.
title Weight-Space Linear Recurrent Neural Networks
topic Machine Learning
url https://arxiv.org/abs/2506.01153