让元学习更稳健，提升系统辨识在极端情况下的可靠性
Distributionally robust minimization in meta-learning for system identification
- 采用分布鲁棒优化，优先处理高损失任务
- 在分布外测试中显著降低安全关键场景的失败率
- 适合对可靠性要求高的系统辨识应用
元学习旨在学习如何解决任务，从而快速适应新场景。现有方法通常优化期望损失，忽略任务间的变异性。本文提出在元学习中采用分布鲁棒优化，优先关注高损失任务，以提升最坏情况下的性能。在一类合成动态系统上训练的元模型，在分布内与分布外设置下均表现优异，显著减少安全关键应用中的失败情况。
原文摘要 · Abstract (English)
Meta learning aims at learning how to solve tasks, and thus it allows to estimate models that can be quickly adapted to new scenarios. This work explores distributionally robust minimization in meta learning for system identification. Standard meta learning approaches optimize the expected loss, overlooking task variability. We use an alternative approach, adopting a distributionally robust optimization paradigm that prioritizes high-loss tasks, enhancing performance in worst-case scenarios. Evaluated on a meta model trained on a class of synthetic dynamical systems and tested in both in-distribution and out-of-distribution settings, the proposed approach allows to reduce failures in safety-critical applications.
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。