发明名称 Displaying examples from texts in dictionaries
摘要 In one embodiment, the invention provides a method for a system to provide information based on a query, the method comprising: performing a first search of at least one first source for information responsive to the query; providing a result of the search to a user; searching documents using at least a part of the result of the search; providing the user with at least one example of usage of the result of the search obtained from the searching of stored documents; based on user input, performing a second search of at least one second source for information responsive to the query; and providing a result of said second search to the user. The invention provides ways of showing the most relevant examples from parallel text corpora according to a ranking.
申请公布号 US9081765(B2) 申请公布日期 2015.07.14
申请号 US201213710060 申请日期 2012.12.10
申请人 ABBYY InfoPoisk LLC 发明人 Kagan Dmitry;Parfentieva Tatiana;Zuev Konstantin;Skopintsev Dmitry
分类号 G06F17/30;G06F17/28;G06F17/27 主分类号 G06F17/30
代理机构 代理人 Weiland LeighAnn;Weinstein Veronica
主权项 1. An electronic device for identifying a set of parallel texts from a corpora of texts, the device comprising: a display capable of displaying a user interface; a processor; a memory in electronic communication with the processor, the memory configured with instructions for performing steps including: identifying a corpora containing two or more parallel texts;receiving a selection of a text fragment in a first language;performing a query on the corpora based on the text fragment to identify a set of parallel texts that include the text fragment in the first language and a substantially equivalent translation of the text fragment in a second language;wherein the identifying the set of parallel texts comprises performing a syntactic and semantic analysis of the text fragments; the syntactic and semantic analysis of the fragments comprises building a language-independent semantic structure for one or more sentence of the text fragments; andestablishing a correspondence between the language-independent semantic structures built for said parallel texts;determining a ranking for the identified parallel texts based on the text fragments, wherein the ranking is based on a frequency of the text fragments occurring in the set of parallel texts; anddisplaying on the user interface the ranked set of parallel texts.
地址 RU