arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1504.07550cs.LGstat.ML
收藏

用于结构化输出预测的深度神经网络正则化

Deep Neural Networks Regularization for Structured Output Prediction

Normandie Univ(诺曼底大学);UNIROUEN(鲁昂大学);UNIHAVRE(勒阿弗尔大学);INSA Rouen(鲁昂国立应用科学学院);LITIS(信息处理与系统实验室)

Soufiane Belharbi, Romain Hérault, Clément Chatelain, Sébastien Adam

摘要:

深度神经网络模型是学习表示的强大框架。通常,它用于通过利用输入$x$中的查询来学习关系$x \to y$。在结构化输出预测问题中,$y$是一个多维问题,各维之间往往存在结构关系。这项工作的动机是学习可能存在于输出数据中的输出依赖性,以提高预测精度。不幸的是,前馈网络无法利用输出之间的关系。为了克服这个问题,我们在本文中提出了一个正则化方案,用于训练神经网络,这些特定的任务,使用多任务框架。我们的计划的目的是在学习监督映射函数$x \to y$的同时,以无监督的方式将输出表示$y$的学习纳入训练过程。我们评估我们的框架上的面部地标检测问题,这是一个典型的结构化输出任务。我们在两个公开的具有挑战性的数据集(LFPW和HELEN)上证明,我们的正则化方案提高了深度神经网络的泛化能力,并加速了它们的训练。未标记的数据和标签的数据的使用也进行了探讨,显示了额外的改善的结果。我们提供了我们框架的开源实现(https://github.com/sbelharbi/structured-output-ae)。

英文摘要:

A deep neural network model is a powerful framework for learning representations. Usually, it is used to learn the relation $x \to y$ by exploiting the regularities in the input $x$. In structured output prediction problems, $y$ is multi-dimensional and structural relations often exist between the dimensions. The motivation of this work is to learn the output dependencies that may lie in the output data in order to improve the prediction accuracy. Unfortunately, feedforward networks are unable to exploit the relations between the outputs. In order to overcome this issue, we propose in this paper a regularization scheme for training neural networks for these particular tasks using a multi-task framework. Our scheme aims at incorporating the learning of the output representation $y$ in the training process in an unsupervised fashion while learning the supervised mapping function $x \to y$. We evaluate our framework on a facial landmark detection problem which is a typical structured output task. We show over two public challenging datasets (LFPW and HELEN) that our regularization scheme improves the generalization of deep neural networks and accelerates their training. The use of unlabeled data and label-only data is also explored, showing an additional improvement of the results. We provide an opensource implementation (https://github.com/sbelharbi/structured-output-ae) of our framework.

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。

↑