This site uses cookies.
Some of these cookies are essential to the operation of the site,
while others help to improve your experience by providing insights into how the site is being used.
For more information, please see the ProZ.com privacy policy.
Kirti Vashee is VP of Enterprise Translation sales for Asia Online. He is a seasoned IT sales and marketing executive and statistical machine translation (SMT) enthusiast who was previously responsible for the worldwide business development strategy at Language Weaver.
He has long-term software industry experience (EMC, Legato, Dow Jones, Lotus) and has been involved in building and managing sales and support operations in Europe and Asia for several software companies.
He is the moderator of the Automated Language Translation group with ovber 4,000 members in LinkedIn and is active on Twitter (@kvashee) and the blogosphere on MT (kv-emptypages.blogspot.com) and translation related issues. He received his formal education in South Africa, India and the United States.
Attila is an ATA-certified English to Hungarian translator, who holds a PhD in physics, and specializes in scientific and technical fields. He is a regular conference speaker who also mentors career-starting translators, putting special emphasis on the intelligent use of diverse tools.
Mirko Plitt is Senior Manager, Language Technologies, in the localization department of Autodesk, the leading maker of CAD software best known for its AutoCAD product family.
In 2001, Mirko led one of the industry’s first deployments of customer-facing raw machine translation, to translate Autodesk’s vast product support knowledge base. He later oversaw the introduction of a company-wide authoring, publishing and translation management system used by hundreds of technical writers, localization engineers and translators. Mirko was then responsible for the integration of machine translation into Autodesk’s translation productivity environment, allegedly the industry’s largest production instance of the open-source Moses system. In this context he organized several rigorous productivity and usability tests which led to unprecedented findings on the real usefulness of machine translation.
Mirko has presented at numerous localization and machine translation events in Europe, the USA and Asia. Most recently, he also organized the first Swiss Translation Unconference, which brought together translators and representatives from industry, government, international organizations, tool makers, education and research.
Don DePalma has more than 30 years of experience in the fields of technology, language services, and market research. As the company’s original founder, he is responsible for launching and developing the preeminent market research firm in the language services sector. He initiated Common Sense Advisory’s coverage of localization maturity, enterprise language processing, business-driven globalization, practical machine translation, return on investment for localization, and multicultural domestic marketing.
As the company's chief strategist, Don serves as an incubator for new ideas and projects that support the organization's vision and growth. He drives strategic decisions affecting the future direction of the research and consulting services. And, he is involved in shaping the strategies of many of the world's largest technology firms and the industry's most influential language services providers via large scale consulting projects.
Prior to founding Common Sense Advisory, he co-founded Interbase Software, served as vice president of corporate strategy at translation technology supplier Idiom Technologies, and was one of the first analysts at Forrester Research, where he consulted to senior management at Global 2000 companies. While at Forrester, Don launched the firm's coverage of various sectors, including content management, application development for strategic internet systems, digital marketing technologies and customer relationship management, ethnic marketing, knowledge management, and business globalization. He lectures, writes, and is frequently quoted on the topics of online marketing, content management, multicultural marketing, localization, return on investment, and website globalization. His book, Business without Borders, is widely used at university and business training courses.
Don holds a PhD in Linguistics from Brown University with specializations in generative grammar, computational linguistics, and the historical phonology of Slavic languages. He has also studied at Moscow State University and Moscow Linguistic University in Russia, Univerzita Karlova in the Czech Republic, and ELISA in Costa Rica. As a linguist, Don studied a range of Indo-European languages and has visited 40 countries.
Dr. Jiri Stejskal is the President & CEO of CETRA Language Solutions. Headquartered in Philadelphia and with an office in Washington, CETRA serves the U.S. Federal Government and corporate clients around the world. Jiri is the immediate past President of the American Translators Association and the current Vice President of the International Federation of Translators (Fédération Internationale des Traducteurs, or FIT).
You must log in before you can access most features of this event.
Discussion for ProZ.com & TAUS present: The Great Translation Debate session (2011): Translation automation is good for the translation profession
Adriana Adarve (X) United States Local time: 18:50 English to Spanish + ...
Machine will never be able to replace humans
Sep 29, 2011
Very interesting topic, but professional image is in jeopardy with the use of MT. A machine will never be able to replace a human, his/her professional, cultural touch, etc., and even less his/her intuitiveness. One of the things mentioned several times was the speed of translation; so, what is more important speed or quality? Another thing mentioned several times was monetary worries. So, what is more important, make tons of money no matter what or actually better communication with all human b... See more
Very interesting topic, but professional image is in jeopardy with the use of MT. A machine will never be able to replace a human, his/her professional, cultural touch, etc., and even less his/her intuitiveness. One of the things mentioned several times was the speed of translation; so, what is more important speed or quality? Another thing mentioned several times was monetary worries. So, what is more important, make tons of money no matter what or actually better communication with all human beings upholding the qualities and the value of languages?
I for one will never be able to agree with machine translations. It is definitely a No win-No win situation. ▲ Collapse
Subject:
Comment:
The contents of this post will automatically be included in the ticket generated. Please add any additional comments or explanation (optional)
The use MT as productivity tool should not lead to lower quality. IMO the idea that post-editing is a completely different skill and should imply making only certain corrections but not others corresponds to the way MT technology was in the past. Unless used "raw" it should be there to help the translator produce high quality translations in shorter time.
Subject:
Comment:
The contents of this post will automatically be included in the ticket generated. Please add any additional comments or explanation (optional)
Attila Piróth France Local time: 00:50 Member English to Hungarian + ...
Agree with Mirko...
Sep 29, 2011
on that MT should be there to help the translator produce high quality translations in shorter time. When the translator is in full control of the MT environment, it can help improve their productivity - just like other tools (e.g., a CAT).
As far as actual usage of this tool is concerned, the parallel with CATs can be drawn further. While most translators agree that translation memories that contain only their own work usually improve their productivity and do not change their... See more
on that MT should be there to help the translator produce high quality translations in shorter time. When the translator is in full control of the MT environment, it can help improve their productivity - just like other tools (e.g., a CAT).
As far as actual usage of this tool is concerned, the parallel with CATs can be drawn further. While most translators agree that translation memories that contain only their own work usually improve their productivity and do not change their normal working methods, translators are often sent projects with large TMs created from many different translators' work. Whether this is a useful tool depends on the quality of the TM. "Well the TM may not be perfect but it was approved by the client, so please do not change it unless you encounter very serious blunders". In this case the TM is no longer the useful tool for the translator it is supposed to be, as you either say no to this assignment, arguing that you can put your skills to better use in projects where mediocre quality is not acceptable - or agree to be much more tolerant to errors.
When MT is used in a way that the first output is out of the control of the translator, the situation is identical, in my opinion. Am I expected to produce as good a work as I can? If not, it means that there are many more people who can also do it. Light postediting falls definitely into this category. High-profile translators are in a way overqualified for it - and they may lack some technical skills that would help them be more productive. Their skill set is not ideally suited to the task. Should they learn the new skills and abandon the old ones? Why not if it is a more lucrative field? But it is not.
In a commoditized setup where big companies own the language data - because develop a sufficiently similarly powerful MT solution is beyond the reach of a single translator - AND quality expectations are lowered, translators become very simply interchangeable. This is nothing new: the same situation already exists with CAT tools. High interchangeability increases the price pressure. So, if pro translators can avoid working in such a competitive market, they should head for their own niche.
Of course, this is not the direct consequence of the tool itself but the way it is used. I perfectly agree with most of the other panelists on that in other scenarios MT can be a very useful aide to the translator. Jeff Allen kindly sent me very detailed comments on his MT experience. By mastering and controlling the technology, he says he could keep high quality and increase productivity very significantly. Thumbs up! He also helped put MT to great use during the Haiti earthquake. Congratulations, it is fantastic stuff.
However, that's not the only way MT is used today. Who owns the tool and who is in control of the data is a crucial question. And a lot of translators who come into contact with MT and PE are in a completely different situation than Jeff. They don't own the tool and they don't control the data. They have language skills - and they may be required to make use of them only partially. And then what will determine their earning potential? Not by the service they provide, because it is a commodity, offered by many. Not by the value they add. But their productivity - and even that not in absolute terms but also compared to their peers.
This is the kind of scenario that most of the respondents to my survey find unattractive. If they can afford, they will move to a different market segment. If...
All this applies only to translators who have access to another market, where their high-quality services are thought. For many newcomers, this market does not open easily, and they may find the possibility of MT PE or other MT-related linguistic tasks a good career option. My fellow panelists can most probably give a better overview of such professions and the great potential they have; as the volume of translation processed by MT is expected to rise considerably ion the future, there are certainly great opportunities in this growing segment of the language-transfer industry.
Machines are already replacing us. By themselves, the replacement is not complete, but provided slave translators post-edit that MT product, the combination of machine (increasingly better) PLUS translator is perfect, a great business for those interested in selling this scheme. And there is something much more imporant (not mentioned in the debate):
By feeding the machine, we are teaching her/him (not to be sexist), and we are going to be less and less necessary.
Machines are already replacing us. By themselves, the replacement is not complete, but provided slave translators post-edit that MT product, the combination of machine (increasingly better) PLUS translator is perfect, a great business for those interested in selling this scheme. And there is something much more imporant (not mentioned in the debate):
By feeding the machine, we are teaching her/him (not to be sexist), and we are going to be less and less necessary.
Saying that MT is an opportunity can be understood from the vendors' point of view, as it IS an opportunity for them. For us, translators, it's bad news.
I am in favor of PE if the software is owned/administered by individual translator as 1) nobody has been left without an assignment job (which is what happens everytime a machine translates, and which has ethical considerations that nobody is analyzing) 2) nobody is contributing to a huge corpus whose intention is to continually train the machine, which, in turn, will mean that fewer post editors (formerly, translators) will be necessary/justified.
This is a perfect plan for the vendors. This is a nightmare for the profession.
what is more important, make tons of money no matter what or actually better communication with all human beings upholding the qualities and the value of languages?
One of the big new trends that global companies are finding is the impact of the web and social networks in particular on the way business is done.
Brian Solis has a new book coming out that describes the overall change quite succinctly. The End of Business As Usual (his new book) expl
what is more important, make tons of money no matter what or actually better communication with all human beings upholding the qualities and the value of languages?
One of the big new trends that global companies are finding is the impact of the web and social networks in particular on the way business is done.
Brian Solis has a new book coming out that describes the overall change quite succinctly. The End of Business As Usual (his new book) explores each layer of the complex consumer revolution that is changing the future of business, media, and culture. As consumers further connect with one another, a vast and efficient information network takes shape and begins to steer experiences, decisions, and markets. It is nothing short of disruptive.
For many companies in the IT and high technology arena this means that much more content needs to be translated to stay in touch with customers who are discussing products and services. Thus more support information needs to be made available and also it becomes important for companies to understand what these social network conversations are saying about their products. Thus this is where you are seeing the most aggressive use of MT. Look at the TAUS membership and the board and you can see who is most motivated to get MT working.
This expanding content often has a high-value for a very short time - maybe weeks or months at most. Thus new production processes are necessary e.g. when Microsoft introduces Windows 7 they need to put a lot of support information about how to deal with issues that are specific to this -- this becomes much less important 6 months later but is critical in the early days as the product is rolled out around the world.
The contents of this post will automatically be included in the ticket generated. Please add any additional comments or explanation (optional)
igna50 Australia Local time: 10:50 English to Spanish
The case for the negative
Sep 29, 2011
It was a most interesting debate but I don’t think the case for the negative was put strongly enough. As the vote shows roughly half of the profession does not agree with the motion. In theory machines will take care of repetitive stuff to let us concentrate on the creative, but in practice what machines bring is “Trados discounts” and a lower price per word. The extra words per hour we may gain do not really compensate for this. The proof: we all know of experienced and talented colleagu... See more
It was a most interesting debate but I don’t think the case for the negative was put strongly enough. As the vote shows roughly half of the profession does not agree with the motion. In theory machines will take care of repetitive stuff to let us concentrate on the creative, but in practice what machines bring is “Trados discounts” and a lower price per word. The extra words per hour we may gain do not really compensate for this. The proof: we all know of experienced and talented colleagues who have left or are thinking of leaving a profession they loved.
You must log in before you can access most features of this event.
