Robust, Scalable Detection of Text Containment in Large Web-Crawled Corpora

Robust, Scalable Detection of Text Containment in Large Web-Crawled Corpora

大规模网络爬取语料库中稳健且可扩展的文本包含检测

Abstract: We present FindMyText, an open-source Python package designed to efficiently assess whether a given text appears, in part or in full, within a text corpus. 摘要: 我们推出了 FindMyText,这是一个开源 Python 软件包,旨在高效评估给定文本是否部分或全部出现在某个文本语料库中。

The tool builds on prior techniques for document fingerprinting, but extends them with a novel mechanism to explicitly capture sequences of matching fingerprints. 该工具建立在现有的文档指纹识别技术基础之上,并通过一种新颖的机制进行了扩展,能够显式地捕获匹配指纹的序列。

By identifying such chains, the tool can more reliably detect near-verbatim copies of a given text rather than mere textual similarities. 通过识别这些链条,该工具能够更可靠地检测给定文本的近乎逐字的副本,而不仅仅是文本相似性。

This makes FindMyText particularly suited for verifying the presence of copyrighted material in a corpus. 这使得 FindMyText 特别适用于验证语料库中是否存在受版权保护的材料。

Leveraging a distributed, disk-based indexing framework, the system scales to large web-crawled datasets. 该系统利用基于磁盘的分布式索引框架,能够扩展至大规模的网络爬取数据集。

Using a new benchmark for evaluating text containment methods, we show that FindMyText outperforms alternative approaches across three datasets (ArXiv papers, Wikipedia, and generic web content). 通过使用一种评估文本包含方法的新基准,我们证明了 FindMyText 在三个数据集(ArXiv 论文、维基百科和通用网络内容)上的表现均优于其他替代方法。


Paper Details:

  • Authors: Lars Henry Berge Olsen, Pierre Lison, Martin Jullum, Mark Anderson
  • arXiv ID: 2607.10020
  • Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
  • Submission Date: 10 Jul 2026

论文详情:

  • 作者: Lars Henry Berge Olsen, Pierre Lison, Martin Jullum, Mark Anderson
  • arXiv ID: 2607.10020
  • 学科: 计算与语言 (cs.CL);人工智能 (cs.AI)
  • 提交日期: 2026年7月10日