cs.LGJun 22, 2026

Fast and Slow Variational Continual Learning

Authors: Subarnaduti PaulYohan JungMohammad Emtiyaz KhanSiddharth SwaroopThomas MöllenhoffMartin Mundt

Organizations: Faculty of Mathematics and Computer Science, University of Bremen, Bremen, Germany · Department of Computer Science & AI, Jeonbuk National University, Republic of Korea · Department of Computer Science, TU Darmstadt, Darmstadt, Germany · RIKEN Center for Advanced Intelligence Project, Tokyo, Japan · Hessian Center for AI (hessian.AI), Darmstadt, Germany · University College London, United Kingdom

Abstract

Continual learning remains a major challenge for modern deep networks, partly because commonly used optimizers lack inherent mechanisms for continual adaptation. One such natural mechanism is fast and slow adaptation to balance stability and plasticity. This mechanism has deep roots in neuroscience and biology, but there is no consensus on how to best incorporate it in commonly used optimizers. Here, we show that this can be easily done via the VCL framework, where past posteriors are used as priors in the future. Our key idea is to incorporate slow adaptation via merging of past posteriors to slow down the drift in the knowledge as learning progresses. The merged posterior is then used as the prior in the VCL update to implement the fast-weight updates. These steps can be seamlessly implemented in the IVON optimizer, whose form and costs are nearly identical to that of Adam. We call this new optimizer the Continual IVON (CoVON) optimizer and show that it not only consistently improves over existing VCL optimizers, but also performs better than other weight-regularization strategies across domain-incremental learning, continual pre-training, and fine-tuning of large language models.

Explore similar work

CardsList
  1. Muon Can Outperform Dedicated Continual Learning Methods

    Sep 21, 2026Sebastian George Sincari, Bogdan Alexandru Gheorghe, Antonio BarbalauEfficient Continual Learning