arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1910.01727cs.LGstat.ML
收藏

广义内环元学习

Generalized Inner Loop Meta-Learning

Facebook AI Research(Facebook人工智能研究院);NYU(纽约大学);University of Southern California(南加州大学)

Edward Grefenstette, Brandon Amos, Denis Yarats, Phu Mon Htut, Artem Molchanov, Franziska Meier, Douwe Kiela, Kyunghyun Cho, Soumith Chintala

摘要:

许多(但不是所有)方法在深度学习和强化学习中自定义为“元学习”,符合近似嵌套优化问题的解决方案的常见模式。在本文中,我们给出了一个形式化的共享模式,我们称之为GIMLI,证明其一般要求,并推导出一个通用的算法实现类似的方法。基于这种分析和算法,我们描述了我们的设计,更高,我们与社区共享,以协助和使未来的研究这些类型的元学习方法的库。最后,我们展示了这个框架和图书馆的实际应用,通过说明性的实验和消融研究,他们促进。

英文摘要:

Many (but not all) approaches self-qualifying as "meta-learning" in deep learning and reinforcement learning fit a common pattern of approximating the solution to a nested optimization problem. In this paper, we give a formalization of this shared pattern, which we call GIMLI, prove its general requirements, and derive a general-purpose algorithm for implementing similar approaches. Based on this analysis and algorithm, we describe a library of our design, higher, which we share with the community to assist and enable future research into these kinds of meta-learning approaches. We end the paper by showcasing the practical applications of this framework and library through illustrative experiments and ablation studies which they facilitate.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑