ICLR 2026 Orals

FIRE: Frobenius-Isometry Reinitialization for Balancing the Stability–Plasticity Tradeoff

Isaac Han, Sangyeon Park, Seungwon Oh, Donghu Kim, Hojoon Lee, KyungJoong Kim

LLMs & Reasoning Sat, Apr 25 · 11:42 AM–11:52 AM · 203 A/B Avg rating: 6.00 (6–6)
Author-provided TL;DR

We present FIRE, a principled reinitialization approach that balances stability and plasticity through constrained optimization.

Abstract

Deep neural networks trained on nonstationary data must balance stability (i.e., retaining prior knowledge) and plasticity (i.e., adapting to new tasks). Standard reinitialization methods, which reinitialize weights toward their original values, are widely used but difficult to tune: conservative reinitializations fail to restore plasticity, while aggressive ones erase useful knowledge. We propose FIRE, a principled reinitialization method that explicitly balances the stability–plasticity tradeoff. FIRE quantifies stability through Squared Frobenius Error (SFE), measuring proximity to past weights, and plasticity through Deviation from Isometry (DfI), reflecting weight isotropy. The reinitialization point is obtained by solving a constrained optimization problem, minimizing SFE subject to DfI being zero, which is efficiently approximated by Newton–Schulz iteration. FIRE is evaluated on continual visual learning (CIFAR-10 with ResNet-18), language modeling (OpenWebText with GPT-0.1B), and reinforcement learning (HumanoidBench with SAC and Atari games with DQN). Across all domains, FIRE consistently outperforms both naive training without intervention and standard reinitialization methods, demonstrating effective balancing of the stability–plasticity tradeoff.

One-sentence summary·Auto-generated by claude-haiku-4-5-20251001(?)

FIRE balances stability-plasticity tradeoff using Frobenius error and isometry deviation constraints without heavy hyperparameter tuning.

Contributions·Auto-generated by claude-haiku-4-5-20251001(?)
  • Proposes principled reinitialization method quantifying stability through Squared Frobenius Error measuring proximity to past weights
  • Quantifies plasticity through Deviation from Isometry reflecting weight isotropy
  • Solves constrained optimization to find reinitialization point minimizing SFE subject to DfI being zero
Methods used·Auto-generated by claude-haiku-4-5-20251001(?)
  • Constrained optimization
  • Newton-Schulz iteration
  • Continual learning
Datasets used·Auto-generated by claude-haiku-4-5-20251001(?)
  • CIFAR-10
  • OpenWebText
  • HumanoidBench
  • Atari
Limitations (author-stated)·Auto-generated by claude-haiku-4-5-20251001(?)
  • Assumes access to past data; does not evaluate FIRE under restricted data scenarios
    from the paper
  • Only evaluated on relatively small models for continual pretraining of LLMs
    from the paper
Future work (author-stated)·Auto-generated by claude-haiku-4-5-20251001(?)
  • Examine FIRE under restricted access to past data
    from the paper
  • Evaluate FIRE on larger models and apply to continual fine-tuning of LLMs
    from the paper

Author keywords

  • stability-plasticity tradeoff
  • continual learning

Related orals

Something off? Let us know →