A new tool for businesses dealing with the issue of multilingual communications was launched this week from a company called SDL. The SDL Automated Translation Solutions tool attempts to solve the language barrier problem by providing instant translations of web content, Microsoft Office documents, instant messages, and emails. It also allows for integration of automated translation into corporate intranet infrastructures and business applications. Has the global language barrier just been broken?
Last month, a post by Marissa Mayer on the Google Blog pointed out the necessity of machine translation for the future of search, saying that the idea of machine-assisted translation is "an incredibly empowering idea" that could "change the way users experience the web and communicate with each other."
That same concept of empowering communications has been incorporated into the new SDL Automated Translations tool, too. The difference with SDL's tool is that instead of just focusing on translations for the web, it also translates documents, emails, chats, company intranet sites, and even internal business applications. Thanks to the tool's open nature, it can be incorporated into anything from customer-facing content on the web to an internal wiki or blog.
The quality of translations can be adjusted to fit your needs, too. For example, you may want your homepage to offer perfect translations of your text, but would rather have on-the-fly, instant translations for use in IM and email. For those quick translations, the tool simply gives users an approximate understanding of sentences and phrases by using something the company calls 'gist' translations.
According to Gilbane Group analyst Leonor Ciarlone, technology advancements and pure computing power have made machine translation not only viable, but also potentially game-changing. A global economy, the volume and velocity of content required to run a global business, and customer expectations is steadily shifting enterprise postures from "not an option" to "help me understand where MT fits."
In their group's Multilingual Communications as a Business Imperative report, they discovered that participants in the study, content management practitioners in multinational organizations, identified machine translation as one of the top three most valuable technologies for the future. Also of note is global communications company Language Weaver's prediction of a potential $67.5 billion market for digital translation, fueled by machine translation. That predication takes into account how new technologies now provide translation at dramatically lowered costs than before. This opens up new, untapped markets, asserts Language Weaver CEO Mark Tapling.
Markets and making money are obviously the focus for the companies involved in these ventures, but we're excited to see machine translation going beyond Google Translate and opening up the business world, too.
Comments
Subscribe to comments for this post OR Subscribe to comments for all Read/WriteWeb posts
Systran has been offering Enteprise Machine Translation solutions for years. In comparison, this new product from SDL doesn't sound like anything new at all.
http://www.systransoft.com/translation/translation-products/server/systran-enterprise-server
Posted by: Bret | October 16, 2008 8:14 AM
@Bret: Hmmm, I thought the option to switch between the 'gisted' translations and human-assisted (that is, high-quality) translations was the difference. I believe Systran's solution is more of an automated offering with a reliance on dictionaries. Please fill me in on what you know.
Posted by: Sarah Perez
|
October 16, 2008 10:59 AM
Bret's right. SYSTRAN has been offering two options for server side (instant) translation (SYSTRANBox and SYSTRANLink) for quite a while. The first of these is the technology powering Google Translator, Babblefish, etc, but you'll find it plugged in behind the scenes all over the place.
How does SDL change anything?
Posted by: Aaron Mentele | October 16, 2008 11:05 AM
There are a number of companies that allow human assisted translations on top of library translation, in real-time or otherwise. GLTaC is one that comes to mind.
Posted by: Aaron Mentele | October 16, 2008 11:08 AM
@Aaron, but GLTaC appears to be a service for requesting translations, not a tool you can integrate with your existing business apps, email, and IM.
Posted by: Sarah Perez
|
October 16, 2008 11:15 AM
I agree with the other comments here, The SDL announcement is not newsworthy and Systran has been doing this for decades already at much higher quality. Google and Microsoft also offer excellent options for gisting quality (much better than SDL) that can be easily determined by doing a comparitive test on the same content.
SDL is basically a company that provides human translation services (with a thin technology veneer)and I think it would not be too harsh to say that the press release was merely an attempt to harvest the growing interest in automated translation to generate more human translation business.
The new interest in MT (after 50 years of failure) is mostly driven by the possibilities being seen by combining SMT (Statistical Machine Translation)with crowdsourcing. Both areas which SDL has no presence in whatsoever.
MT coupled with large scale human feedback can enable systems to improve at a rate that we have not seen yet. Since MT can produce large amounts of content filled with linguistic errors, it is possible to clean this up if a crowd of capable/competent humans can be motivated to help. The popular term for this is phenomena is “crowdsourcing”. We have seen this at work on a small scale, at Facebook already, and at Asia Online we are embarking on a 3 million page translation of the English Wikipedia into Thai initially, then into several other Asian languages. This approach will be used to translate tens of millions of pages and gradually raise this content to human quality levels with assistance from the broad student community that would find the content most useful.
http://news.tourthailand.org/business-news/online-proofreaders-sought.html
http://www.bangkokpost.com/100908_Database/10Sep2008_data62.php
MT together with web based massive online collaboration is emerging as a model that can take on huge translation tasks and we see now several initiatives around the world beginning to explore this model. What is special about these efforts is that we are seeing is actually a social phenomenon coming to a focus around a collaborative technological platform involving machine translation.
Alain Desilets of the NRC of Canada recently said, “"Two technologies which will drastically change the way we translate content: massive online collaboration a la Wikipedia, and Machine Translation. Shared language data repositories are central to both the collaborative and MT innovations. A year ago, I would have said that MT was still too imperfect to impact the translation industry in any significant way. But recently, progress has been incredibly rapid, even more rapid than its most optimistic proponents ever dreamt of."
http://www.wiki-translation.com/tiki-index.php?page=Processes+and+tools+for+massively+collaborative+translation
Brian McConnell of the Worldwide Lexicon Blog makes a prediction in an interesting article on this site:
“The language barrier, as we know it, will be gone by 2010. Computer scientists have been chasing a Holy Grail of machine intelligence for decades, but the breakthrough that will eliminate the language barrier is social, not technical. Language, like music or art, demands people to comprehend it.”
He goes on to say,
“The language barrier will be broken down in a series of simple steps. The first phase of this transition will be driven by publishers with large or highly motivated audiences. These early adopters will recognize the value of making their content visible in many languages, and their readers will be happy to contribute. Each website will develop its own translation community from its audience. At this stage of the transition, the system will be driven by a few publishers, and probably a few thousand dedicated translators.
As these projects grow, and as multilingual publishing tools become more sophisticated, aggregators will emerge. These sites will create large translation communities that decide what to translate based on their interests, whether or not a particular publisher is aware of this activity. Roaming mobs of amateur translators will translate whatever they think is interesting. Commercial services that complement volunteer based systems will also appear.”
The full article can be seen at http://blog.dermundo.com/original/2356.html
In terms of the market potential, I think it is clear that it is growing but both Gilbane and Language Weaver need to apply more facts to stay credible. One more clarification to get the facts straight and let readers decide for themselves on what is plausible and realistic.
There are two markets that we are talking about and it is important to get some clarity on what they are:
1. The Overall Translation Services Market
2. The Actual Direct Market for MT Technology
The Overall Market:
Common Sense Advisory estimates (after much care and research, I presume) that the overall market for 2008, human-delivered translation activities “will total a hefty US$14.25 billion” growing to US$24B by 2012.
http://www.commonsenseadvisory.com/members/res_cgi.php/080528_QT_2008_top_25_lsps.php
Now it can be argued that this is just the tip of the iceberg and that the real potential is much higher. So if you make the gross assumption, that buyers will pay exactly the same amount for what they do NOT translate, as they do for what they do translate, you can come up with some basic estimates.
A standard iceberg is 90% submerged so if we apply the above assumption, we can say the market could grow to $240B by 2012.
If we say that the tip of the iceberg is only 5% of what could be translated, you could even get this forecast number up to $480B by 2012. (Wouldn’t that be nice!). If we go further and apply the same assumption to the 99.44% of digital content that CSA says is not translated, we can come up with numbers that are in the very high trillions. Even Carl Sagan would be humbled.
Unfortunately as we all know (and as Sarah Palin is discovering), we do have to apply reason to these gross assumptions. The primary reason that this content is not being translated, is that it does not make economic sense to do so at the same rates paid for the tip of the iceberg. If costs could be lowered while maintaining a level of quality, people probably would translate more content. This is something that CSA could research and perhaps suggest some sort of relationship between, cost/quality/content. That would be useful to us all and also help to identify the core value relationships between different types of previously un-translated content, quality required and viable cost scenarios.
The Actual Direct Market for MT Technology:
In the same article referenced in the CSA blog entry, CSA says, “On the software side, we estimate that the MT software market falls well short of US$100 million.” Actually, if you read previous research produced by CSA they estimate this number to be more like $50M in 2007 http://www.globalwatchtower.com/2007/12/20/mt-eyeballs/ .
So if you consider that the Language Weaver announcement claims that the “digital translation market” (i.e. non human) will grow from $50M to a potential $67.5B by 2011, I think we need to ask, how? Or as they say in SNL, OH REALLY !?!!
We as an industry lose credibility when others, initially take this information seriously and subsequently discover the real underlying facts. We already see people like Gilbane take this at face value http://gilbane.com/globalization/. If these analysts don’t question the veracity of press releases, it makes sense for us as the community to raise the questions. We all stand to gain from clear well founded information and we all stand to lose from wrong information. I am also engaged with a company that produces SMT technology solutions, and would love to have this $67B market come true. I would be a direct beneficiary. However, the history of MT is filled with eMpTy promises. If this technology is to rise above its historic reputation of being crap/ barely usable, we are going to need a better strategy than announcing outlandishly large market potential numbers to drive growth.
As I have said earlier in this discussion, the growth potential for MT is really great, though (I think) unlikely to have a monetized value of $67.5B in 5 years. Microsoft and Google built their businesses and the overall industry by consistently delivering value to an expanding set of customers. This is probably the path that we in MT will also have to take.
I believe in substantial growth for MT in the coming future, but I think we will need changes to drive this growth. There are two things that, I think, can drive the MT technology market into high gear:
1. Clearly better quality
2. Much greater and deeper human collaboration (Massive Online Collaboration)
At Asia Online we have also made an aggressive and perhaps overly optimistic forecast. One that rises above the expert CSA forecast. We believe this is possible by creating better quality with much more intensive human collaboration. The NIST competition results show that the raw technology will evolve at a very slow pace and we need something more to really kick this into high gear. I believe that one of the keys to market success will be to really get the language service providers on board to drive the high end high quality market.
For the record here is the Asia Online forecast for what could happen if we get the growth drivers into place:
http://www.gizmag.com/asia-online--the-worlds-most-significant-literacy-project/10049/gallery/picture/53176/
Finally, the one thing we all have in common is that we are all optimistic about the future of automated translation and that is a good thing as that is probably a necessary condition to make change and real growth happen. At the end of the day, forecasts do not make or break markets and some may even say, that they don’t even really give you a true indication of anything other than what we know today and what we wish might happen.
Posted by: Kirti | October 16, 2008 11:35 AM
Hi Sarah. They have services not covered well on their site. I can't speak at much length for GLTaC, though.
What I can say is that (human / library) hybrid translation is not new - we've plugged in similar solutions for email correspondence as many as 3 years ago using third-party solutions. And SYSTRAN has allowed instant translation for email, etc for a long time.
I wouldn't be surprised if SDL is using SYSTRAN for the instant translation and then a team of human translators as editors. This is what other services do.
I hope that SDL's aim is lessening implementation cost and/or improving real-time translations. Otherwise, I still question what's new or improved.
Posted by: Aaron Mentele | October 16, 2008 1:06 PM
Thanks commenters for filling me in. Esp @Kirti - wow. A blog post in a comment!! Excellent. Lots of info to research now.
Posted by: Sarah Perez
|
October 16, 2008 2:04 PM
There is no better machine translator site than: http://www.windowslivetranslator.com/
try it!
Posted by: Murilo D'Agguilar | October 16, 2008 5:52 PM
A key use case of Systran's Enterprise solution is to integrate Machine Translation into the localization process, with the end goal of publishing high quality translations (human translator level). You can find a very interseting case study here:
http://www.systransoft.com/translation/systran/corporate-profile/translation-case-studies/symantec
Achieving a high level of quality requires an investment in customization of the MT engine to be tuned to a certain domain or corpus. Using Systran this traditionally has meant building custom dictionaries, as it is a rule based MT system. However, Systran is now starting to offer hybrid statistical/rule based engines, which means the customization can also include statistical language models. In a case like Symantec, who does post editing of MT output, this means that the system can automatically "learn" from the corrections made by human translators. As commenter Kirti says, it is an exciting time for MT technology.
So yes, the notion of switching between gisting and high quality modes has always been an important part of Systran's offering. Quality comes with an investment and it is a graded scale. A customer like Symantec can have different quality requirements for different types of translation work: online technical support pages can have a lower quality than product documentation, as the value is in the ability to immediatly push out time sensitive information in the case of technical support.
Posted by: Bret | October 17, 2008 3:21 AM
Gilbane focuses on content technologies and their application to high-value business solutions. We apply quantitative and qualitative research to our discussions of current and expected enterprise challenges. Our Multilingual Communications as a Business Imperative report provides the enterprise view of expectations for integrated and *interoperable* content technologies that help manage content volume and velocity, meet global customer expectations, and raise the visibility of multilingual communications to a C-level priority. Gilbane does not benchmark performance, rate feature sets and capabilities of solutions, or produce market forecasts. Having said that, we cover and comment on industry announcements that can have an impact on improving what we call the Global Content Value Chain.
Posted by: Leonor Ciarlone | October 17, 2008 10:54 AM
GLTaC provides a mix of Human and Machine Translation services, which optimizes the end result for the customer. The ultimate goal of any translation software or service is to take something from one language to another, quickly and conveniently. The quality level of the final translation and speed at which it occurs are ultimately the customer's decision.
Since we use a variety of MT software (not all done by one package) we seek "Best of Breed" MT capability.
Posted by: Doug Strock | November 6, 2008 8:14 AM