PATENTSCOPE Search Service Adds New Multilingual Search Functionality

WIPO’s PATENTSCOPE search system has recently added Cross-Lingual Information Retrieval (CLIR) in beta mode for public testing within its PCT and National Patent Data Collections search system. This interface allows you to search the PCT database of International Applications, as well as the patent collections of ARIPO, Cuba, Israel, Korea, Mexico, Singapore, South Africa and Vietnam. When you use the CLIR tool, your search query in one language will be translated into several other languages using a number of cross-lingual dictionaries built in-house by WIPO. Currently, cross-lingual search is available in the following languages:

  • English
  • French
  • German 
  • Japanese 
  • Spanish

Begin by entering a search string in the search box and selecting the language of your query (the default search language is English). Next, select from two choices under the “Expansion Mode”. AUTOMATIC will simply generate the results immediately without any further user input.  For more control and a better view of how the system works, select SUPERVISED.

01_initial_query

You are also able to decide if precision or recall are more important in your search. Precision is a measure of exactness and will retrieve fewer documents which should be more targeted to your query. Recall is a measure of completeness which will retrieve a larger number of results, some of which may be less on-point to your query; higher recall represents a “scorched earth” type of search.

In the Supervised mode, you will next be provided with a list of technical domains intended to focus your search to the relevant technological field. The system will propose likely domains associated with your query in the right column, but you can de-select the suggested domain and/or add more domains from the list. Choosing the areas that are most closely related to your search topic will help the program select the most appropriate cross-lingual synonyms from its dictionaries.

02_technical_domains

Next, the system will suggest synonyms to add to your search. Simply select the variants that you determine are relevant. You can also increase or decrease the number of synonyms the system will propose by moving the More or Less slider. Finally, you can manually add synonyms that are not in the proposed list. Click on TRANSLATE SELECTED TERMS to add the new variants to your query.

03_variants

The final search screen allows you to decide which fields to search, the distance allowed between the search words, and whether or not to enable word stemming. Word stemming allows the system to search for variations on the root form of the word; for example, if you search “swim”, the results will include swimming, swimmers, etc. Before deciding to search in the Description or Claims fields only, please be sure to refer to the database’s coverage information to determine whether or not this would be an effective strategy; full text publications are not available for all of the patenting authorities in the system.

04_fields

05_proximity

Clicking on one of the language tabs at the top of the search box will automatically show you the translated search phrase.

06_language

In addition, there is a tab for International Patent Classification codes in which the system generates a basic set of main IPC classes to be included in the search string. These class codes are joined into the search query using the AND operator, helping to ensure that the search is covering the appropriate general technology area.

07_IPC

Closing a language tab by clicking on the X will remove that translated search string from the query.

The search results are presented with various sorting options, as well as a machine translation function which uses Google Translate to provide titles and abstracts from foreign patents in the language of your choice.

09_translate

Basic analysis tools are also available that allow you to view the top assignees, patenting offices, inventors, and IPC codes represented in your result set. Clicking on any of the hyperlinks in the analysis feature will automatically re-focus your result set further.

08_analysis

While it is still in Beta, this is a powerful search system and a promising entry into the patent search market. Its inclusion of various national collections can provide access to bibliographic data and PDF documents that are not available elsewhere. In particular, the inclusion of the Mexican and South African information has provided otherwise hard-to-obtain documents in the past. Best of all, it’s free to use.

A step-by-step user guide on CLIR is available on the WIPO web site. If you find yourself on WIPO’s PCT and National Patent Data Collections search page without linking directly to the CLIR interface, you can switch easily to CLIR by hovering over the “Search” tab at the top of the page and selecting “Cross Lingual Expansion” from the drop-down menu.

10_CLIR