To examine the architectural determinants of robustness, perturbations that alter connectivity topology, excitatory/inhibitory sign structure, weight magnitudes, and weight placement are constructed while preserving complementary properties.
Abstract
Reservoir computing provides a controlled setting for studying how recurrent network architectureshapes computation: input signals are projected into a high-dimensional state space by a fixed nonlinear dynamical system, and only the readout is trained. However, reservoir performance can be dependent on hyperparameters; this paper asks which recurrent network features support robustness to those parameter changes. We characterize computational performance using memory capacity (MC), truncated single-delay information-processing capacity (IPC), and kernel rank (KR). Generalization across input histories is measured using generalization rank (GR), while hyperparameter robustness is quantified using the coefficient of variation (CV) of each metric across sweeps of target spectral radius, input scaling, leak rate, and neuron bias. To examine the architectural determinants of robustness, we construct perturbations that alter connectivity topology, excitatory/inhibitory sign structure, weight magnitudes, and weight placement while preserving complementary properties. Across these experiments, the C. elegans connectome consistently occupies a relatively low-variance regime. The central result is a performance-robustness tradeoff: architecture variants with higher task-agnostic performance also tend to exhibit greater hyperparameter sensitivity and poorer common-tail generalization. Across the E/I edge balance sweeps and shuffle controls, this tradeoff is closely associated with the raw spectral radius before normalization. Because every perturbed matrix is rescaled to the same target radius, matrices with lower raw spectral radius receive greater global amplification of their recurrent weights. The observed differences among architectures therefore characterize the joint effects of structural variation and architecture-specific global rescaling under spectral-radius normalization.
This work introduces a deterministic, pilot-informed selector for leaky linear reservoirs followed by coordinate-wise nonlinear features and positions free-probability kernels as deterministic surrogates for selecting reservoir operating regimes when validation rollouts are scarce.
Sara Malacarne, Andrea Ceni, Claudio Gallicchio· 0 citations
The results establish neuronal cultures as state-dependent living reservoirs whose native dynamics support computation and are incorporated into a closed perception–action loop for Mario Kart control using a fixed readout.
Samitha S. Somathilaka, Jacob Clouse, Sasitharan Balasubramaniam· bioRxiv· 0 citations
In computational neuroscience, task-trained recurrent neural networks (RNNs) are commonly used as a testbed for exploring the space of recurrent circuit solutions compatible with a particular function. However, these networks are subject to inductive biases that may be misaligned with those of biological circuits, whic...
Recurrent neural networks (RNNs) offer linear-time scaling with sequence length while requiring only constant memory, yet they struggle to capture long-range dependencies due to vanishing gradients and limited receptive fields. To address these limitations, we introduce a second-order recurrent model in which the stand...
Aleix Salvador-Pomarol, A. Montanari, E. K. Miller et al.· 0 citations
Training RNNs on delayed decision-making tasks with progressively increasing temporal demands shows that temporal and decision-related computations can emerge through multiple dynamical regimes, while maintaining structured low-dimensional representations and comparable behavioural performance, mirroring biological pri...
C. Jarne, Ryeongkyung Yoon, Tahra L. Eissa et al.· Journal of Computational Neu...· 0 citations
This work systematically characterize a large family of local plasticity rules that were meta-learned in prior work, and shows that their performance is predictable and structured: high-scoring rules are characterized by strong differentiation between neurons encoding the ball trajectory and the background, with stable...
Maciej Kania, Basile Confavreux, T. Vogels· bioRxiv· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 20, 2026
Senior Research Scientist Jim Glass discusses how his MIT-IBM Computing Research Lab collaboration connects perception and language, and addresses the shortcoming of current models.