公开(公告)号：WO2018076058A1

公开(公告)日：2018-05-03

申请号：PCT/AU2017/051175

申请日：2017-10-26

Applicant: COMMONWEALTH SCIENTIFIC AND INDUSTRIAL RESEARCH ORGANISATION

Inventor： POSTNIECE, Linda ,  HANLEN, Leif ,  SIMON, Travis ,  BACON, Neil ,  GOVERNATORI, Guido

IPC: G06F17/20 ,  G06F17/21 ,  G06F17/27 ,  G06N5/02 ,  G06F17/30 ,  G09B7/08

Abstract: This disclosure relates to computationally efficient processing of legal texts. A language processor extracts from text blocks atomic literals that correspond to terms in the legal text and that are shared across logic expressions for different text blocks. A mapper maps the text blocks into a logic expression of the atomic literals and a logic engine evaluates the logic expression based on one or more assignments of the atomic literals. A user interface comprises user input elements associated with the atomic literals. Upon user interaction with the user input elements, an assignment is created of the atomic literal associated with the user input elements and the logic engine is called to evaluate the logic expression based on the created assignment. The user interface changes based on the evaluation to thereby provide legal advice in the form of remaining atomic literals.

Abstract translation: 本公开涉及对法律文本进行计算有效的处理。 语言处理器从文本块中提取与法律文本中的术语相对应的原子文字，并且在不同文本块的逻辑表达式之间共享。 映射器将文本块映射为原子文本的逻辑表达式，逻辑引擎基于原子文本的一个或多个赋值来评估逻辑表达式。 用户界面包括与原子文字相关联的用户输入元素。 在用户与用户输入元素交互之后，创建与用户输入元素相关联的原子文字的分配，并且调用逻辑引擎以基于创建的分配评估逻辑表达式。 用户界面基于评估而改变，从而以剩余原子文字的形式提供法律建议。

公开(公告)号：WO2018102861A1

公开(公告)日：2018-06-14

申请号：PCT/AU2017/051315

申请日：2017-11-29

Applicant: COMMONWEALTH SCIENTIFIC AND INDUSTRIAL RESEARCH ORGANISATION

Inventor： HANLEN, Leif ,  BACON, Neil ,  NOCK, Richard

IPC: G06F19/24 ,  G06Q50/22

Abstract: This disclosure relates to automatic text analysis. A source device comprises an input port to receive text data representing text that comprises multiple text segments. A text processor transforms the text data to numeric values that represent the multiple text segments. An encryption engine encrypts the numeric values into encrypted numeric values. An aggregation device aggregates the encrypted numeric values into aggregated encrypted numeric. Finally, a learning device remote from the source device provides an analysis of the text data by performing a learning method on the aggregated encrypted numeric values using operations that maintain the aggregated encrypted numeric values and thereby the text data concealed from the learning device. The numeric data values allow encryption to preserve privacy while, at the same time, allowing learning on the numeric data on a remote device without disclosing the text data because the aggregated numeric values remain concealed from the learning device.