The Intelligent Automation of Contract Analysis of Collateral Warranties

抵押担保合同分析的智能自动化

• 批准号: 104899
• 金额: $ 35.55万
• 依托单位: BEALE & COMPANY SOLICITORS LLP
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2019
• 资助国家: 英国
• 起止时间: 2019 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=104899
• 关键词: Intelligent; Automation; Contract; Analysis; Collateral

"Beale & Company are construction industry lawyers. We spend a great deal of time reviewing contracts on behalf of our clients, one particular type of contract known as a Collateral Warranty is a particular bugbear.The work in reviewing Collateral Warranties is an important part of commercial risk management, can be legally complicated but is often undervalued and under-resourced. If it could be automated, even partially, all concerned would benefit.Current contract review software is largely machine learning-based and cannot achieve the detailed legal analysis required and how it achieves its results is often opaque. On the other hand, rule-based approaches are overly complex and frail.Beale & Company have teamed up with the University of Swansea to develop prototype contract review software to overcome these shortcomings.It will combine both machine learning and rule-based approaches in an innovative way. We will harness the high level review power of machine learning to identify manageable parts of contracts and the close precision of rules to analyse the legal content those parts in detail. The rules will be simple and robust.It will be developed for lawyers by lawyers.The software will be based on the open source natural language processing software, 'GATE' developed by the University of Sheffield.While we initially develop a prototype for CWs, we believe we can then extend the approach to other types of construction contracts and to other contracts generally."

“Beale & Company是建筑行业的律师。我们花费大量时间代表客户审查合同，其中一种特殊类型的合同被称为抵押担保，这是一个特别的问题。审查抵押担保的工作是商业风险管理的重要组成部分，可能在法律上很复杂，但往往被低估和资源不足。如果能够实现自动化，即使是部分自动化，所有相关方都将受益。目前的合同审查软件主要是基于机器学习的，无法实现所需的详细法律的分析，而且它如何实现其结果往往是不透明的。另一方面，基于规则的方法过于复杂和脆弱。Beale & Company与斯旺西大学合作开发了原型合同审查软件，以克服这些缺点。它将以创新的方式联合收割机机器学习和基于规则的方法。我们将利用机器学习的高水平审查能力来识别合同的可管理部分，并利用规则的精确性来详细分析这些部分的法律的内容。这些规则将是简单而稳健的。它将由律师为律师开发。该软件将基于谢菲尔德大学开发的开放源代码自然语言处理软件“GATE”。虽然我们最初为CW开发了一个原型，但我们相信我们可以将这种方法扩展到其他类型的建筑合同和其他合同。"