arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.09253stat.MLcs.LG
收藏

用于结构化预测的深度草图输出核回归

Deep Sketched Output Kernel Regression for Structured Prediction

LTCI, Télécom Paris(电信巴黎理工学院 LTCI 实验室);IP Paris(巴黎理工学院);University of Milan(米兰大学)

Tamim El Ahmad, Junjie Yang, Pierre Laforgue, Florence d'Alché-Buc

AI总结:

本文提出深度草图输出核回归,用随机近似核协方差算子的特征函数构造有限维输出子空间,使神经网络可借助核诱导损失和梯度下降完成结构化预测。

摘要:

通过在输出空间中利用核技巧,核诱导损失为定义适用于多种输出模态的结构化输出预测任务提供了一种有原则的方法。特别是,它们已在代理非参数回归场景中得到成功应用,在该场景中核技巧通常也被用于输入空间。然而,当输入是图像或文本时,深度神经网络等表达能力更强的模型似乎比非参数方法更合适。在本工作中,我们探讨如何训练神经网络来解决结构化输出预测任务,同时仍受益于核诱导损失的多功能性和相关性。我们设计了一族新颖的深度神经架构,其最后一层在由核诱导损失导出的无限维输出特征空间的一个数据相关有限维子空间中进行预测。该子空间被选为经验核协方差算子的随机近似版本的特征函数张成的空间。有趣的是,该方法解锁了梯度下降算法(进而解锁任意神经架构)在结构化预测中的使用。在合成任务以及真实世界监督图预测问题上的实验证明了我们方法的相关性。

英文摘要:

By leveraging the kernel trick in the output space, kernel-induced losses provide a principled way to define structured output prediction tasks for a wide variety of output modalities. In particular, they have been successfully used in the context of surrogate non-parametric regression, where the kernel trick is typically exploited in the input space as well. However, when inputs are images or texts, more expressive models such as deep neural networks seem more suited than non-parametric methods. In this work, we tackle the question of how to train neural networks to solve structured output prediction tasks, while still benefiting from the versatility and relevance of kernel-induced losses. We design a novel family of deep neural architectures, whose last layer predicts in a data-dependent finite-dimensional subspace of the infinite-dimensional output feature space deriving from the kernel-induced loss. This subspace is chosen as the span of the eigenfunctions of a randomly-approximated version of the empirical kernel covariance operator. Interestingly, this approach unlocks the use of gradient descent algorithms (and consequently of any neural architecture) for structured prediction. Experiments on synthetic tasks as well as real-world supervised graph prediction problems show the relevance of our method.

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。

↑