Report:PatBase/Viewing Results/Viewing Full Text/Machine Translations for Non-Latin Text
From Intellogist
| Report | Patent Coverage Map | Ratings | Comments |
| This search system report was created by the Intellogist Team and is available for viewing only. If you'd like to share your knowledge on Intellogist, please visit the Best Practices, Glossary, or Community Reports pages. If you are a registered user and would like to be notified of any substantial changes to this report, you may place a "watch" on the Revisions page, which is the last page listed on the table of contents. To learn more about using the Intellogist "watchlist," see the Watchlist Help page. | 
|
  |
|
Machine Translations for Non-Latin Text
Non-Latin text records are now searchable in PatBase through both the command line and the Non-Latin Text Search form. However, once users find and view their non-Latin text search hits, they may wish to read them in English or another language via a machine translation tool.
A machine translation tool for Latin text records is discussed in an earlier section of this article. However, both the PatBase internal translation tool and the Google translation tool do not function correctly for all non-Latin text languages. As of December 2011, not every non-Latin language in PatBase can be translated. The following table shows which non-Latin text records currently have a machine translation option available in PatBase (outside translator tools can still be used for those languages that do not have this option in-system): [1]
| Authority | Apparent language of publication available in PatBase | On-Demand Google Machine Translations or PatBase Internal Translation Available? | Notes |
| Bulgaria (BG) | Bulgarian | On-Demand Google Machine Translation - Yes On-Demand PatBase Internal Translation - No |
Google translation only accessible through "Translate" option in View/Browse hit list. English-language machine translated abstracts present for some records. |
| China (CN) | Chinese | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes On-Demand PatBase Internal Translation (through both Hit List and Full Text) - Yes |
Pre-loaded searchable machine translations for Chinese patent applications, utility model applications, and utility models available. |
| Eurasian Patent Office (EA) | Russian | On-Demand Google Machine Translation - Yes On-Demand PatBase Internal Translation - No |
Google translation only accessible through "Translate" option in View/Browse hit list. English-language machine translated abstracts present for some records. |
| Greece (GR) | Greek | On-Demand Google Machine Translation - Yes On-Demand PatBase Internal Translation - No |
Google translation only accessible through "Translate" option in View/Browse hit list. English-language machine translated abstracts present for some records. |
| Japan (JP) | Japanese | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes On-Demand PatBase Internal Translation (through both Hit List and Full Text) - Yes |
Pre-loaded searchable machine translations for Japanese patent applications available from January 1998 to September 2011. Some available from 1992. |
| Korea (KR) | Korean | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes On-Demand PatBase Internal Translation (through both Hit List and Full Text) - Yes |
|
| Morocco (MA) | Arabic | On-Demand Google Machine Translation - No On-Demand PatBase Internal Translation - No |
No Arabic language records available as of December 2011. Titles available in French. |
| Patent Cooperation Treaty (PCT) | Japanese, Chinese, Russian, potentially Arabic (unconfirmed) | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes (for Chinese, Japanese, and Russian) On-Demand PatBase Internal Translation (through both Hit List and Full Text) - Yes (for Chinese and Japanese) |
The PatBase internal translation Translate option seems to display "Chinese" as the default starting language for these records, even when the records are Japanese-language or Russian-language original text. This can lead to erroneous translations unless the user selects the appropriate language from the drop-down menu before beginning. For the Russian-language PCT documents, a translation option via the internal PatBase translation is offered from Russian to English in the full-text view, but the translation was unintelligible when tested. |
| Russia (RU) | Russian | On-Demand Google Machine Translation - Yes (View/Browse hit list only) On-Demand PatBase Internal Translation - No |
Google translation only accessible through "Translate" option in View/Browse hit list. Coverage of Russian language non-Latin text seems to extend beyond only the PCT collection. English-language machine translated abstracts present for some records. |
| Serbia and Montenegro (YU) | Serbian | On-Demand Google Machine Translation - No On-Demand PatBase Internal Translation - No |
No original language records available as of December 2, 2011. Some records have English-language titles. |
| Taiwan (TW) | Chinese | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes On-Demand PatBase Internal Translation (through both Hit List and Full Text) - Yes |
|
| Thailand (TH) | Thai | On-Demand Google Machine Translation (through both Hit List and Full Text) - Yes On-Demand PatBase Internal Translation - No |
Google translation accessible through both Full Text and "Translate" option in View/Browse hit list. English-language machine translated abstracts present for some records. |
| Turkey (TR) | Turkish | On-Demand Google Machine Translation (through Hit List only) - Yes On-Demand PatBase Internal Translation - No |
The drop-down menu in the Google translation menu from the Hit List option doesn't list Turkish in the language drop down menu. The user must first translate the document to view the second available Google Translate menu above the translated text, where the option for Turkish is now available. It may just be easier to cut and paste the full non-Latin text directly into Google Translate in a separate browser. |
If a translation tool is available for a non-Latin text record in PatBase, this option will display below the document number when viewing the full record. A translation option is also available from both the View and Browse mode hit lists, where a "Translate" link is displayed in the bar above each record that lists the family number. The Google Translate service and internal PatBase translation service are used to produce machine translations within PatBase. Google Translate is a free web tool which has been made available directly from the PatBase user interface. (It is also available from this site on the web.)
The figure below shows a Chinese non-Latin text record for which machine translation is available. (Note that PatBase highlighting features do not apply to non-Latin text records.)
Alternatively, when no translation options are available through the full-text view, the non-Latin text will display minus those options. As an example, the figure below shows a Bulgarian record for which no translation features are available in the full-text view (the Translate option on the family record in the View/Browse hit list will allow the user to translate the document using Google Translate).
Editor's Note:Although the limitations of current machine translation technology are well documented, it is worthwhile to note here that even when machine translation is available, the results can be as mystifying as if they were still in another language. Note that the highlighting options that are available in the machine translator for Latin text will not work on the non-Latin text, making it difficult to understand where search keywords are represented in the translation.
It may also be difficult for users to navigate to the available translation options, since for some languages, the translation option is only accessible through the Translate option above the family record in the hit list.
Sources
- ↑ Availability of translation options was determined by empirical testing, conducted in PatBase on December 2, 2011.
