Registry and Payments dataset

The dataset used in this paper is a binary problem, a pair of sentences can be either similar or not similar, with a high density of personal information and identification codes, and the sentences don't always have a clear meaning but mostly represent stacked information.

本文所采用的数据集为二元分类任务数据集，任务目标为判定给定句子对是否相似。该数据集包含高密度的个人信息与标识码，且句段未必具备清晰语义，多数情况下仅承载堆叠式信息。