提出新方法让神经网络更高效生成权重，适合频繁更新的场景。
Learning an Efficient Optimizer via Hybrid-Policy Sub-Trajectory Balance
- 分两阶段生成权重，解耦优化策略提升灵活性。
- 结合在线与离线学习，提升局部优化能力，避免长时推理问题。
- 在少样本、迁移学习等任务中表现更优，推理更快。
生成式建模的进展使得神经网络可在无需梯度优化的情况下生成权重。然而，现有方法存在过度耦合与长时程问题：前者将权重生成与特定任务目标紧密绑定，限制了优化器的灵活性；后者因缺乏局部约束，导致推理效率低且准确率差。本文提出 Lo-Hp，一种解耦的两阶段权重生成框架，通过学习多种优化策略增强灵活性。其采用混合策略子轨迹平衡目标，融合在线与离线学习以捕捉局部优化策略。理论上证明，仅学习局部优化策略即可缓解长时程问题，并提升全局最优权重的生成能力。实验验证，Lo-Hp 在需频繁更新权重的任务（如迁移学习、少样本学习、域泛化及大语言模型适配）中表现出更高的精度与推理效率。
原文摘要 · Abstract (English)
Recent advances in generative modeling enable neural networks to generate weights without relying on gradient-based optimization. However, current methods are limited by issues of over-coupling and long-horizon. The former tightly binds weight generation with task-specific objectives, thereby limiting the flexibility of the learned optimizer. The latter leads to inefficiency and low accuracy during inference, caused by the lack of local constraints. In this paper, we propose Lo-Hp, a decoupled two-stage weight generation framework that enhances flexibility through learning various optimization policies. It adopts a hybrid-policy sub-trajectory balance objective, which integrates on-policy and off-policy learning to capture local optimization policies. Theoretically, we demonstrate that learning solely local optimization policies can address the long-horizon issue while enhancing the generation of global optimal weights. In addition, we validate Lo-Hp's superior accuracy and inference efficiency in tasks that require frequent weight updates, such as transfer learning, few-shot learning, domain generalization, and large language model adaptation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。