Grammar A

该数据集通过语法A生成了多种方程，用于训练和测试MGMT及其表格数据集嵌入架构的能力。这些方程包括最高至4次的多项式、双变量的三角函数、反函数、对数函数和指数函数。该数据集是基于语法规则生成的，这些规则对语法树的大小和常量的范围设有限制。数据集的规模根据生成的方程而变化，其任务是发现方程和表格数据集的嵌入。

This dataset generates diverse equations via Grammar A for training and testing the capabilities of MGMT and its tabular dataset embedding architecture. These equations include polynomials up to the 4th degree, bivariate trigonometric functions, inverse functions, logarithmic functions, and exponential functions. This dataset is generated based on syntactic rules that impose restrictions on the size of syntax trees and the range of constants. The size of the dataset varies with the generated equations, and its target task is to learn embeddings for both equations and tabular datasets.