As in any other task, the process of building machine learning models can
benefit from prior experience. Meta-learning for classifier selection gains
knowledge from characteristics of different datasets and/or previous
performance of machine learning techniques to make better decisions for the
current modeling process. Meta-learning approaches first collect meta-data that
describe this prior experience and then use it as input for an algorithm
selection model. In this paper, however, we propose an automatic learning
scheme in which we train convolutional networks directly with the information
of tabular datasets for binary classification. The goal of this study is to
learn the inherent structure of the data without identifying meta-features.
Experiments with simulated datasets show that the proposed approach achieves
nearly perfect performance in identifying linear and nonlinear patterns,
outperforming the traditional two-step method based on meta-features. The
proposed method is then applied to real-world datasets, making suggestions
about the best classifiers that can be considered based on the structure of the
data.

本文介绍一种基于元学习的分类器自动选择算法，通过使用卷积神经网络直接学习表格数据的内在结构，来提高二元分类的性能，并在模拟和真实数据集上实验证明，这种方法优于基于元特征的传统两步方法。

卷积神经网络分类中的自动学习算法选择

Automatic learning algorithm selection for classification via  convolutional neural networks

Much of the work in metalearning has focused on classifier selection,
combined more recently with hyperparameter optimization, with little concern
for data preprocessing. Yet, it is generally well accepted that machine
learning applications require not only model building, but also data
preprocessing. In other words, practical solutions consist of pipelines of
machine learning operators rather than single algorithms. Interestingly, our
experiments suggest that, on average, data preprocessing hinders accuracy,
while the best performing pipelines do actually make use of preprocessors.
Here, we conduct an extensive empirical study over a wide range of learning
algorithms and preprocessors, and use metalearning to determine when one should
make use of preprocessors in ML pipeline design.

本文通过 metalearning 方法研究了机器学习中预处理数据的作用以及何时应该运用预处理器在 ML 管道设计中发挥作用，实验发现预处理器可以提高 ML 管道的性能。