NeuMoSync: End‑to‑End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning

ICLR 2026 Conference Submission20716 Authors

Keywords: Continual Learning, Neuroscience-inspired AI, Loss of Plasticity, Knowledge Transfer, Neuromodulation

TL;DR: This paper introduces NeuMoSync, a novel deep neural network architecture inspired by brain neuromodulation that enhances plasticity and adaptability in continual learning by integrating dynamic, neuron-specific modulation.

Abstract: Continual learning (CL) requires models to learn tasks sequentially, yet deep neural networks often suffer from plasticity loss and poor knowledge transfer, which can impede their long-term adaptability. Drawing inspiration from global neuromodulatory mechanisms in the brain, we introduce $\textbf{Neu}$ro$\textbf{Mo}$dulation and $\textbf{Sync}$hronization ($\texttt{NeuMoSync}$), a novel architecture that integrates dynamic, neuron-specific modulation into deep neural networks to enhance their adaptability and plasticity. $\texttt{NeuMoSync}$ extends standard neural network architectures with learnable feature vectors per neuron that tracks network-wide historical context and incorporates a module operating at a higher level of abstraction. This module synthesizes neuron-specific signals, conditioned on both current inputs and the network’s evolving state, to adaptively regulate activation dynamics and synaptic plasticity. Evaluated on diverse CL benchmarks, including memorization (Random Label CIFAR-10, Random Label MNIST), concept drift (Shuffle CIFAR-10), class-incremental (Class Split T-ImageNet, CIFAR-100) and domain-incremental (Permuted MNIST), $\texttt{NeuMoSync}$ demonstrates strong performance in terms of retention of plasticity and achieves improvements in both forward and backward adaptation compared to existing methods. Ablation studies validate the necessity of each component, while analysis of the learned modulatory signals reveals emergent behaviors that parallel neuroscience observations. Our work underscores the potential of integrating brain-inspired global coordination mechanisms into deep learning systems to advance robust, adaptive continual learning.

Primary Area: transfer learning, meta learning, and lifelong learning

Submission Number: 20716