Noradrenergic-inspired gain modulation attenuates the stability gap in joint training

Alejandro Rodriguez-Garcia, Anindya Ghosh, Srikanth Ramaswamy

Keywords: stability gap, continual learning, neuromodulation, stability-plasticity dilemma

TL;DR: We mitigate the stability gap in continual learning with a noradrenaline-inspired gain mechanism that adapts learning rates and flattens the loss landscape, reducing interference across tasks.

Abstract: Recent work in continual learning has highlighted the *stability gap* -- a temporary performance drop on previously learned tasks when new ones are introduced. This phenomenon reflects a mismatch between rapid adaptation and strong retention at task boundaries, underscoring the need for optimization mechanisms that balance plasticity and stability over abrupt distribution changes. While optimizers such as momentum-SGD and Adam introduce implicit multi-timescale behavior, they still exhibit pronounced stability gaps. Importantly, these gaps persist even under ideal joint training, making it crucial to study them in this setting to isolate their causes from other sources of forgetting. Motivated by how noradrenergic bursts transiently increase neuronal gain under uncertainty, we introduce a dynamic gain scaling mechanism as a two-timescale optimization technique that balances adaptation and retention by modulating effective learning rates and flattening the local landscape through an effective reparameterization. Across domain- and class-incremental MNIST, CIFAR, and mini-ImageNet benchmarks under task-agnostic joint training, dynamic gain scaling effectively attenuates stability gaps while maintaining competitive accuracy, improving robustness at task transitions.

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

Submission Number: 12882