Intellectual Property Office

From Saltlux

◈ Project Contents ◈
Customer: Korea Intellectual Property Office
Field: Public/Patent
Name of Project: Existing Technology Auto Comparison System (KIPO)
Period of Project Accomplishment: August 2004년 ~ December 2005
Applied Solution: [IN2]DOR, [IN2]TMS
URL: www.kipo.go.kr


◇ <Patent> becomes a very important constituent element of an enterprise as the competition has become seriously increased. In this social environment the patentees want to protect their proprietary product patent and to will the competition. Accordingly they want to search the way to win the competition For these the patent managers of the industries and the related government authorities will invest a significant amount of various resources including time and money to do search the similar patents.

Current KIPO Situation
KIPO is a government authority who helps industries and individuals to protect their intellectual properties for their contribution of their own knowledge properties to the technical innovation and to the industrial. KIPO controls in all the fields including but not limited to trademark, electricity/electronics, information technology, and bio-technology and so on. Number of patent administration makes decision for the similarities of the applied patent to the existing ones. To do this effectively and efficiently the patent administration are facing difficulties and invest significant investment of time and budget as the have to investigate the foreign and patents together with the domestic ones.



Current issues of KIPO systemic
Existing keyword style patent retrieval system was unable to decide the similarities of the patents as the existing retrieval system only applied simple keyword matching technology and unable to systemically decide the similarities because of the impossible detect of the real meaning of the paten documents. So the patent administration’ needs are automatic analysis of the patent documents, differentiation of the contents of the new document to the existing document, listing-up the documents in order of high similarities and provide to the administration the automatic comparison results against the existing documents. In addition, the administration needs the <semantic based retrieval system> to efficiently process express the retrieval intention through pre-understanding of the contents of the documents at the query creation.

Application of [IN2]DOR과 [IN2]TMS 적용
Intellectual existing technology retrieval system must enable to automatic retrieval base on the feature extraction, classification clustering and summarization functions by understanding the internal contents of document through supporting retrieval functions according to the weight of quality and through high quality language analysis.
[In2]DOR and [In2]TMS were applied for the individual retrieval in the documents and text mining. Document similarities statistical information was created based on the statistics and rule. Language module was constructed to enable natural language process against the user query and the stability and safety was improved in connection with [IN2]DOR as the qualitative and high functional retrieval engine
The limitation of the retrieval was overcome by enabling the retrieval of the contents of the equivalent material relationship which is the most important inspection element through use the use of various knowledge resources including but not limited to thesaurus, technical terminology, new terminology synonym dictionary


Flow of Existing Technology Automatic Comparison Retrieval

In the existing technology auto comparison system the users have been satisfied simultaneously in two ways of <Keyword based query> based on the keyword retrieval technology and <document based query> based on the natural language process under the connection with the PatentNet II. With this system the patent applicant’s application documents will be understood, intellectually analyse and provides the similar cases to the patent administration patent when he inputs so that the administration understands the contents of the technology and reduces the burden of inspection work.

Effect of System Functional Improvement
Indexing speed was improve by 6 times by application of patent textual sentences and Index Cache of the DBMS of KIPO, and the retrieval time was save by 1/5 for indexing a million cases in compare to the existing system. Retrieval functions were improved by un-used terms/similar terms dictionary reinforcement, retrieval speed improvement by Multi-Level Clustering, and by supporting query words extension

Maximization of Patent Examination Work Efficiency
Through construction of <Similar patent integrated search system> KIPO could have improved productivity by minimization of patent examiner’s workload due to possibility of document based search accomplishment by appointment of patent application. Workload on query preparation has also been minimized through NLP based automatic query generation.


Effect of Automatic Comparison System Construction of Advanced Technology

Number of search has been minimized by efficient query generation through displaying search result arrayed in order of similarity between documents and displaying similarities by identification items. As a result, world fastest patent examination process was built together with improvement of business efficiency through application of <intelligent patent analysis system> in examination business of <patent advanced technology differentiation>.