Pre-generated Expressions Dataset

该数据集包含了一百万个预先生成的表达式，这些表达式允许最多一个嵌套操作，并且最多包含五个一元运算符。此外，该数据集旨在分析多变量系统中每个变量产生的影响，并已用于训练回归神经网络。其规模为训练用的一百万个表达式和验证用的一十万个表达式。该数据集的任务是训练一个多集合变换器，以生成一元符号框架。

This dataset comprises one million pre-generated expressions, each allowing at most one nested operation and containing up to five unary operators. Furthermore, this dataset is designed to analyze the impact of each variable in multivariate systems, and has been utilized to train regression neural networks. It consists of one million training expressions and one hundred thousand validation expressions. The task associated with this dataset is to train a multi-set Transformer to generate unary symbolic frameworks.