CLIR launches an attack on the Tower of Babel
April 21, 2011
Searching by keyword in the descriptions and claims of patents written in Chinese, Korean, Russian or Portuguese is now possible in PATENTSCOPE without prior knowledge of these languages.
In June 2010, WIPO launched the beta version of the
CLIR –
Cross-Lingual Information Retrieval - for PATENTSCOPE : an innovative multilingual information
search tool in 5 languages (German, English, Spanish, French and Japanese) which allows patent
material searches to be extended.
Less than a year after launching this
tool,
Chinese, Korean, Russian and Portuguese have been added. Korean, Russian and Portuguese are still
at an experimental stage due to the complexity of the Korean language and the limited size of
training corpuses for Portuguese and Russian.
The following improvements have also been added:
- Improved synonym propositions in Japanese in supervised mode.
- German language coverage has been extended using the patent collection published by the EPO.
In order to improve the coverage of our dictionaries, WIPO is looking for bilingual corpuses
in the following language combinations : English/Portuguese, English/Spanish and
English/Russian. We are particularly looking for parallel texts of titles, abstracts and scientific
articles in these bilingual combinations, which will allow us to considerably improve the
CLIR tool
from and into these languages.
Your opinion and suggestions to improve the tool are
welcome.