arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2001.11578stat.MLcs.LG
收藏

从数据中学习假设空间II:收敛性与可行性

Learning the Hypotheses Space from data Part II: Convergence and Feasibility

Diego Marcondes, Adilson Simonis, Junior Barrera

摘要:

在部分\textit{I}中,我们提出了一个一般假设空间$\mathcal{H}$的结构,学习空间$\mathbb{L}(\mathcal{H})$,它可以用来避免\textit{overfitting}估计时,在一个复杂的空间相对缺乏的例子。此外,我们提出了U曲线的性质,可以利用它来选择一个假设空间,而无需穷尽搜索$\mathbb{L}(\mathcal{H})$。在本文中,我们通过展示基于学习空间的模型选择框架的一致性来进一步推进我们的议程,在该框架中,人们从数据中选择要学习的假设空间。本文开发的方法通过将Vapnik-Chervonenkis理论扩展到\textit{random}假设空间,即,从数据中学习的假设空间。在这个框架中,估计随机子空间$\hat{\mathcal{M}} \in \mathbb{L}(\mathcal{H})$,其以概率1收敛到具有期望属性的目标假设空间$\mathcal{M}^{\star} \in \mathbb{L}(\mathcal{H})$。由于收敛意味着渐近无偏估计量,我们有一个一致的框架模型选择,这表明它是可行的学习假设空间从数据。此外,我们表明,在$\hat{\mathcal{M}}$上学习的泛化错误小于在$\mathcal{H}$上学习时所犯的错误,因此在从数据中学习的子空间上学习更有效。

英文摘要:

In part \textit{I} we proposed a structure for a general Hypotheses Space $\mathcal{H}$, the Learning Space $\mathbb{L}(\mathcal{H})$, which can be employed to avoid \textit{overfitting} when estimating in a complex space with relative shortage of examples. Also, we presented the U-curve property, which can be taken advantage of in order to select a Hypotheses Space without exhaustively searching $\mathbb{L}(\mathcal{H})$. In this paper, we carry further our agenda, by showing the consistency of a model selection framework based on Learning Spaces, in which one selects from data the Hypotheses Space on which to learn. The method developed in this paper adds to the state-of-the-art in model selection, by extending Vapnik-Chervonenkis Theory to \textit{random} Hypotheses Spaces, i.e., Hypotheses Spaces learned from data. In this framework, one estimates a random subspace $\hat{\mathcal{M}} \in \mathbb{L}(\mathcal{H})$ which converges with probability one to a target Hypotheses Space $\mathcal{M}^{\star} \in \mathbb{L}(\mathcal{H})$ with desired properties. As the convergence implies asymptotic unbiased estimators, we have a consistent framework for model selection, showing that it is feasible to learn the Hypotheses Space from data. Furthermore, we show that the generalization errors of learning on $\hat{\mathcal{M}}$ are lesser than those we commit when learning on $\mathcal{H}$, so it is more efficient to learn on a subspace learned from data.

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。

↑