Gained in translation: Google comes to help the EPO

"No wonder it won't work", said Professor Prout,
"one of the French irregular verbs has broken loose"
Now here's something to ponder.  A media release, "European Patent Office and Google sign memorandum on translation of patents" (here), heralds the exciting prospect that the deadlock over European patent linguistic issues might eventually be broken by the adoption of machine translation solutions.  According to the media release, dated 30 November,
"The European Patent Office (EPO) and Google have ... signed a Memorandum of Understanding to improve access to patent translations in multiple languages.

Under the envisaged collaboration, the EPO will use Google's machine translation technology to translate patents into the languages of the 38 countries that it serves. In return, it will provide Google with access to its translated patents, enabling Google to optimise its machine translation technology. Google technology will be used to translate patents originating in Europe as well as patents originating in other regions of the world and enjoying protection in Europe.

The collaboration aims to offer faster and cheaper fit-for-purpose [this begs the question: "which purpose?" It seems to the IPKat that patents are read for more than one reason] translations of patents for companies, inventors and scientists in Europe. Today, anyone wishing to register a patent must do so in one of the EPO's official languages - English, French and German. They then need to arrange for translation of the patent - at their own cost - into the languages of all countries in which they wish the patent to apply. This complexity means that many European patents are not available in all national languages or legally binding in all the EPO's member states. Similarly, anyone searching for information in patents published in foreign languages finds it difficult to retrieve data relevant to their research projects.

"The European Patent Office is one of the largest providers of free information on state-of-the-art technology disclosed in patents from around the globe. The partnership with Google to create machine translation tools for patents will help inventors, engineers and R+D teams to retrieve relevant documents efficiently - in their own language - from our wealth of published patent information. This agreement with Google puts the EPO at the forefront of efforts to strengthen the patent system's international character and improve its quality for the benefit of the global economy," said EPO President Benoît Battistelli.[Merpel is always amused at the way in which matters pertaining to patent law are said to be beneficial to the global economy. One wonders how the global economy ever managed in the past, particularly since patents seem to be irrelevant to the economy in so many parts of the globe]

The collaboration also aims to facilitate the decision process of the EU states in their attempt to simplify the introduction of a single pan-European patent [Does this mean that it's going to be opposed?]. While the EPO provides a common entry point to obtain patents throughout Europe, patent owners must still validate and, where necessary, translate patents in each individual EU country. As a result, obtaining Europe-wide patent protection is significantly more expensive than in markets such as the US. A single EU patent could reduce costs and enhance legal certainty [Please can someone explain how a single EU patent is more legally certain than several national patents? The criteria of patentability and revocation aren't being changed, are they? ], giving businesses and innovators unified protection for their inventions.

For Google, the collaboration offers a major opportunity to improve its translation service. The EPO will offer access to around 1.5 million documents, and each year this number grows by more than 50 000 new patent grants [how does translating patent documents improve Google's translation service more than translating non-patent documents, wonders Merpel]. The partnership also covers Asian languages. Facilitating access to the rapidly increasing volumes of Japanese, Chinese and Korean technological information is one of the biggest challenges facing the global patent system.

"This collaboration is exciting for both Google and Europe. It will help to increase access to information for all Europeans, supporting the innovation process and allowing the European economy to strengthen its competitiveness," said Carlo d'Asaro Biondo, Google's Vice President of Sales for Southern and Eastern Europe, the Middle East and Africa. "It demonstrates how private companies can work with public institutions to find innovative solutions to difficult issues."
The IPKat welcomes anything and everything that will make patent documents more widely accessible. However, he notes that nothing is said about how Google, which is not known for its philanthropic contributions to the wellbeing of international institutions such as patent-granting authorities, is going to monetise this arrangement.  Are translated terms and the interrelationship of those terms with the names of companies and inventors going to be filtered into Adsense software and other means of targeting prospective consumers?

Merpel says, when are they going to invent a device that translates English-language patent documents into English too?

Cat language here
Languages that start with Cat here and here
Gained in translation: Google comes to help the EPO Gained in translation: Google comes to help the EPO Reviewed by Jeremy on Tuesday, December 07, 2010 Rating: 5

10 comments:

  1. This is a fascinating development, and anything that encourages European competitiveness is to be applauded. However, I have grave misgivings as to why, up to now, the EPO has been telling us that its machine translations have been perfectly fit for purpose and, when it decides that they are not, it enters into an agreement with a major player in the patent landscape to "rectify" the shortcomings it previously told us were not there with no public consultation.

    Surely it would have been better to forge links with a translation house with little interest in the patenting process or the Court of Justice translation service?

    Let's have some pretence at due process....

    ReplyDelete
  2. Where is that spirit for the winter. Merpel is a gloomy cat. Of course snow does affect cats like that -- they ostentiatiously flick their feet after experiencing the white stuff.

    On begging the purpose -- my experience is that cats are quite good at begging -- good for any purpose would be better than not having translations at all, but a pretty good starting purpose is that they should be good enough to alert readers to a possible issue on validity or infringement. Maybe better if they would answer the question completely, but wise men/women say that even a professional translation will not necessarily achieve that.

    Global economy -- well in the pre-patent centuries they relied on not telling anyone, or disposing of anyone who knew. Now -- admittedly Western commentators -- say repeatedly when reviews are carried out, that patent protection is fundamental to wide area of technological development. Maybe not to software, but definitely for pharmaceuticals, even if other steps need to be taken to assist poorer countries to be able to use them.

    A single patent more certain: well once invalidated it is invalidated in all countries. No more valid here, invalid there, valid somewhere else again. No more sarcastic comments about different decision in different jurisdictions. Of course whether the one and only decision is right, well???? But is is certain.

    And on how translating patents improves future translating of patents, Merpel answer the question in his last comment. Patent attorneys write a different language (all over the world). Systems need to learn that language. As Merpel says perhaps there is real mileage in a programme which can translate English patents into English.

    My vote is in favour of Google's collaboration.

    ReplyDelete
  3. ...how does translating patent documents improve Google's translation service more than translating non-patent documents...

    To google, the EPO's collection is like a huge cavern-full of Rosetta Stones! Where else will you find such a corpus of documents that (are supposed to) say and mean exactly the same thing in a number of different languages, professionally translated at a high quality level?

    Google's translation technology is based on statistical methods. It is improved with every opportunity to compare similar texts in different languages. This deal is a boon for them, even before they start leveraging it to improve targeting of advertising!

    ReplyDelete
  4. Dear Merpel,
    why only translations from Patentese to English? There is also a strong demand for Patentisch to German and for Brevetais to French. Perhaps even from Octroois to Dutch.

    ReplyDelete
  5. Google doesn't have to monetise the arrangement. They likely consider it worthwhile for "close" access to what must be one of (if not the) largest collections of professionally translated documents in existence.

    The existence of professional human translations of patent documents is surely what "improves Google's translation service more than translating non-patent documents". Their technical nature is not really relevant.

    What matters to Google is improving their auto-translate algorithms. A robust auto-translate system allows them to generally improve their search index and ad-targeting.

    Also, note that the Google translate home page does not serve any ads. The translate service, like many of Google's offerings, is simply a way to improve the depth of Google's index and to keep them at the top of the pile.

    ReplyDelete
  6. Merpel,

    I generally enjoy your interspearsed little comments in the citations at the IPKat, but this time I think you were a bit too bantering.

    Anyone who has tried to get anything useful out of a machine translation of a patent text knows that what comes out is usually complete jibberish. Any improvement of the machine translations will improve its fit-for-purposeness.

    Do you have any particular nation in mind where people live in healty prosperity and where patents are irrelevant to the economy? I have a hard time finding an example.

    A single EU patent would be more legally certain than several national patents at least for the reason that today, a patent can be found valid and infringed in one country, but invalid or not infringed in another.

    Finally, access to one of the world's largest patent databases will at least not limit Google's improvements of the translation service, even less so the improvements relating to translation of patent documents.

    And yes - it would be really nice if someone invented a machine that translates English-language patent documents into proper English too! :-)

    ReplyDelete
  7. Example of automatic translation of a claim into English:

    "An apparatus for atrampart mice, consisting a wood base, means in form of letter Or, and an enlistment with a cheese tip of such form that when the mouse eats the cheese kills it to the apparatus."

    ReplyDelete
  8. @Anonymous 11:09

    How do we know that the original isn't equally nonsensical?

    ReplyDelete
  9. Google patents is really easy for the patent illiterate to search compared with espacenet, but unfortunately is limited to US patents.

    Regardless of the merits of the translation, if Google have the corpus of European and national patents available for search in the same way as they have US patents that will be a very useful alternative route from espacenet for free patent searching.

    ReplyDelete
  10. Hmm, his is quite a change in attitude. Compare with this:
    http://www.intellogist.com/wiki/Report:Google_Patent_Search/Security

    Shortly after the system’s release, the EPO’s Principal Director of Patent Information, Wolfgang Pilch, raised this important point (and it was probably raised by others, as well): no one in the patent community should forget that Google is profitable, as a company, through tracking user activity and analyzing user search patterns. Therefore any search performed on Google is hypothetically vulnerable to scrutiny, and sensitive searches should not be performed on it.

    Google's don't be evil is now accepted by the EPO?

    ReplyDelete

All comments must be moderated by a member of the IPKat team before they appear on the blog. Comments will not be allowed if the contravene the IPKat policy that readers' comments should not be obscene or defamatory; they should not consist of ad hominem attacks on members of the blog team or other comment-posters and they should make a constructive contribution to the discussion of the post on which they purport to comment.

It is also the IPKat policy that comments should not be made completely anonymously, and users should use a consistent name or pseudonym (which should not itself be defamatory or obscene, or that of another real person), either in the "identity" field, or at the beginning of the comment. Current practice is to, however, allow a limited number of comments that contravene this policy, provided that the comment has a high degree of relevance and the comment chain does not become too difficult to follow.

Learn more here: http://ipkitten.blogspot.com/p/want-to-complain.html

Powered by Blogger.