Analyzing Mathematics to Detect Disguised Academic Plagiarism

该项目的目标是改进对变相学术抄袭的检测方法，如转述、翻译和思想抄袭，特别是对科学、技术、工程和数学(STEM)领域的文件。为了实现这一目标，我们将进行基础研究，将数学表达式作为与文本无关和语言无关的特征来分析，以确定文档的潜在可疑相似度。我们将把新的基于数学的剽窃检测方法与我们在先前研究中开发的基于文本和基于引文的检测方法结合起来。考虑不同特征类型进行相似性评估的组合方法最有希望检测范围广泛的学术抄袭形式。为了使基于数学的方法能够检测抄袭，以及在实践中的其他信息检索任务，我们将研究从各种输入格式的科学文档中提取和语义增强数学表达式的方法。在这项研究任务中，项目组将与Bela Gipp的副总干事资助的项目GI 1259/1-1“推进从数字图书馆检索数学知识的方法和工具，用于搜索、推荐和辅助系统”密切合作。该项目开发数学信息检索方法，以满足维基百科内的信息需求。拟议的项目将扩展这些方法，使其适用于科学文献。对于教育和研究机构、资助机构和学术出版商来说，检测学术剽窃是一个紧迫的问题。然而，目前，抄袭检测提供商专注于识别学生中流行的非混淆形式的抄袭。出于经济原因，他们避免了必要的研究和开发努力，以便能够发现更强的模糊形式的剽窃，这是研究论文的特征。为了使研究界和抄袭检测提供商能够更好地识别精心伪装的学术抄袭实例，我们将在开源和免费使用抄袭检测系统HyPlag(www.hyplag.org)中实施我们的贡献，并公开提供我们研究的代码和数据。为了确保我们的研究满足领域专家的需求，我们将在zbMATH(www.zbMATH.org)的专业数学编辑的帮助下，不断评估我们的研究贡献。通过与数学专家的密切合作，我们进一步打算为未来在数学信息检索方面的贡献奠定基础，而不仅仅是抄袭检测用例。

The goal of this project is to improve the detection methods for disguised forms of academic plagiarism, such as paraphrases, translations and idea plagiarism, particularly for documents in the fields of Science, Technology, Engineering and Mathematics (STEM). To achieve this goal, we will conduct fundamental research on analyzing mathematical expressions as text-independent and language-independent features to determine potentially suspicious similarity of documents. We will combine the novel mathematics-based approach to plagiarism detection with text-based and citation-based detection approaches that we developed in previous research. Combined approaches that consider different feature types for similarity assessment are most promising to detect the wide range of academic plagiarism forms. To enable the mathematics-based approach to plagiarism detection, as well as other information retrieval tasks in practice, we will research methods for extracting and semantically augmenting mathematical expressions from scientific documents in various input formats. For this research task, the project team will collaborate closely with the Bela Gipp’s DFG-funded project GI 1259/1-1 “Methods and Tools to Advance the Retrieval of Mathematical Knowledge from Digital Libraries for Search-, Recommendation- and Assistance-Systems”. The project develops mathematical information retrieval methods to serve information needs within the Wikipedia. The proposed project will extend these methods and make them applicable for scientific documents.Detecting academic plagiarism is a pressing problem, e.g., for educational and research institutions, funding agencies, and academic publishers. However, currently, plagiarism detection providers focus on identifying the non-obfuscated forms of plagiarism prevalent among students. For economic reasons, they avoid the research and development effort necessary to enable the detection of more strongly obfuscated forms of plagiarism, which are characteristic of research papers. To enable the research community and plagiarism detection providers to better identify carefully disguised instances of academic plagiarism, we will implement our contributions in the open source and free to use plagiarism detection system HyPlag (www.hyplag.org) and openly provide the code and data of our research. To ensure our research meets the needs of domain-experts, we will continuously evaluate our research contributions with the help of professional mathematics editors from zbMATH (www.zbMATH.org). By collaborating closely with mathematics experts, we furthermore intend this project to lay the groundwork for future contributions in Mathematical Information Retrieval beyond the plagiarism detection use case.