Automatic reasoning over contract clauses

• Ano de defesa: 2019
• Autor(a) principal: Aires, João Paulo de Souza
• Orientador(a): Meneguzzi, Felipe Rech
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Departamento: Escola Politécnica
• País: Brasil
• Palavras-chave em Português: Conflitos normativos; Lógica deôntica; Processamento de linguagem natural; Processamento automático de contratos
• Palavras-chave em Inglês: Norm conflicts; Deontic logic; Natural language processing; Contract process automation.
• Área do conhecimento CNPq: CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/9256

Contracts formally represent agreements between parties and often involve the exchange of goods and services. In contracts, norms define expected behaviours from the parties using deontic statements, such as obligations, permissions, and prohibitions. However, norms may conflict invalidating themselves and producing a contract inconsistency. A conflict often arises when two or more norms are applied to the same context but have different deontic statements, such as permissions x obligations and prohibitions x obligations. The identification and resolution of such conflicts is often made by humans, which makes the task time-consuming and error-prone. In order to automate such identification, in this thesis we introduce an approach to identify and classify norm conflicts between norms in contracts written in natural language. We rely on the use of sentence embeddings to represent and manipulate natural language to extract information and use it to identify norm pairs as conflicts. We propose four norm conflict classes and use them to train a norm conflict classifier that can help on the conflict cause identification. The results show that our approach achieves an accuracy higher than 99% on the identification and 78% on the classification of conflicts.