English

Trust Region Continual Learning as an Implicit Meta-Learner

Machine Learning 2026-05-28 v2

Abstract

Continual learning aims to acquire tasks sequentially without catastrophic forgetting, yet standard strategies face a core tradeoff: regularization-based methods (e.g., EWC) can overconstrain updates when task optima are weakly overlapping, while replay-based methods can retain performance but drift due to imperfect replay. We study a hybrid perspective: \emph{trust region continual learning} that combines generative replay with a Fisher-metric trust region constraint. We show that, under local approximations, the resulting update admits a MAML-style interpretation with a single implicit inner step: replay supplies an old-task gradient signal (query-like), while the Fisher-weighted penalty provides an efficient offline curvature shaping (support-like). This yields an emergent meta-learning property in continual learning: the model becomes an initialization that rapidly \emph{re-converges} to prior task optima after each task transition, without explicitly optimizing a bilevel objective. Empirically, on task-incremental diffusion image generation and continual diffusion-policy control, trust region continual learning achieves the best final performance and retention, and consistently recovers early-task performance faster than EWC, replay, and continual meta-learning baselines.

Keywords

Cite

@article{arxiv.2602.02417,
  title  = {Trust Region Continual Learning as an Implicit Meta-Learner},
  author = {Zekun Wang and Anant Gupta and Christopher J. MacLellan},
  journal= {arXiv preprint arXiv:2602.02417},
  year   = {2026}
}

Comments

21 pages, 21 tables

R2 v1 2026-07-01T09:32:26.645Z