translations-620-1
2 December 2012PatentsIan Pearce

Translations: man v machine

In its December, 2011 edition of the World Intellectual Property Indicators report, the World Intellectual Property Office (WIPO) stated that in 2010 worldwide patent applications reached a record 1.98 million documents.

A large proportion of the recent growth in patenting activity originated in China where, WIPO indicates, between 2001 and 2010 annual growth averaged 22.6 percent, with patent filings rising from 63,450 in 2001 to 391,177 in 2010. While this reflects the rise of China as a worldwide industrial and economic power, it does bring language difficulties for non-native patent researchers.

Of course, language concerns are not limited to Chinese documents. Japan and Korea are also important patent-issuing countries where language barriers may cause difficulties. Even within Europe, language considerations have until recently hindered the progression of a Community or Unitary patent which would grant legal protection in all participating countries.

As patent numbers have risen, one strategy to help overcome language considerations has been the increased use of machine translations (MT) to aid or replace human translations, allowing a patent search to be conducted in English irrespective of the language of the original document.

One major advantage of MT over human translations is the speed at which the translation can be made available. For example, the Chinese patents full-text database from the Lighthouse IP Group, which is currently available on the Dialog platform, processes all State Intellectual Property Office documents (applications, grants and utility models) within one to two weeks of publication using the latest statistical MT techniques. This is far quicker than could be achieved by human translators.

However, expert human translators do supplement aspects of the original document, such as inventor and assignee names, which cannot always be reliably translated by a machine but which are oft en crucial to a successful patent search. Supplementary information such as legal status events is also translated and made available to facilitate effective freedom-to-operate analysis in China.

“IT CAN BE USEFUL TO NARROW THE SCOPE OF RETRIEVAL TO SPECIFIC PARTS OF THE FULL TEXT ONLY, AS IS POSSIBLE ON THE DIALOG PLATFORM.”

Searching any type of full-text document can be tricky, as the increased number of words available for retrieval means that simple Boolean searches using ‘and/or’ are likely to retrieve large numbers of potentially unwanted results. The diligent use of proximity operators is therefore required to ensure that any search terms have the correct relationship to each other. This is even more important when working with MT as there is also the risk of strange sentence structures which have been carried forward from the original language into the English translation.

In addition, even though MT technology has improved considerably in the last decade it is likely to struggle with any nuances in the text requiring more lateral thinking than usual on the searcher’s part when constructing strategies.

Depending on the needs of the search it can be useful to narrow the scope of retrieval to specific parts of the full text only, as is possible on the Dialog platform, such as the claims for freedom-to-operate searches.

Another potentially useful addition for any document which has been translated is the ability to review the original text and associated images. Linking through to an original PDF containing drawings and formulae, for example, can often help clarify any potential ambiguities in the translated text.

The launch of patent content next year on the new ProQuest Dialog platform will mean the Japanese patents full-text database from Lighthouse IP Group also becomes available in combination with additional collections of full-text patent databases from LexisNexis Univentio.

One of the advantages of patent documentation is that an application must be made in every country in which protection is sought and be prepared in the language applicable to that country. Expanding any search into patent family databases such as INPADOC or Derwent World Patents Index for equivalent family applications in a more familiar language can oft en be beneficial.

It should always be noted that there may be differences between the actual texts of individual documents from different countries, but even taking this into account, equivalents can be a useful way to help clarify any deficiencies in a machine-translated document.

Already registered?

Login to your account

To request a FREE 2-week trial subscription, please signup.
NOTE - this can take up to 48hrs to be approved.

Two Weeks Free Trial

For multi-user price options, or to check if your company has an existing subscription that we can add you to for FREE, please email Adrian Tapping at atapping@newtonmedia.co.uk


More on this story

Patents
1 July 2013   The European Patent Office (EPO) has opened up its free translation service for patent documents to enable access for six more languages.