Abstract
We investigate the use of Amazon Mechanical Turk for the creation of translations from English to Haitian Creole. The intention is to produce a bilingual corpus for Statistical Machine Translation. In several experiments we offer varying amounts of money for the translation tasks. The current results show that there is no clear correlation between pay and the translation quality. Almost all translations show a significant overlap with online translation tools which indicates that the workers did often not translate the sentences themselves.
1 Introduction
Our group is currently involved in the development of an English↔Haitian Creole translation system for use in the earthquake region of Haiti. One of the current tasks is the rapid production of a bilingual English↔Haitian Creole in-domain medical dialogue corpus to be able to train a Statistical Machine Translation system. Some native Haitian Creole speakers volunteered to help with translations and we also intend to use professional translators to support the effort.
Amazon’s Mechanical Turk (AMT) is an interesting alternative here as it would be cheaper than using professional translators. This is especially relevant for an English↔Haitian Creole translation system as the commercial potential is probably limited.
One of the main concerns with using AMT for NLP tasks, especially translation is the quality of the resulting data and the availability of workers with Haitian Creole knowledge.
The experiments in this paper address these concerns and evaluate the translations produced by Amazon Mechanical Turk compared with professionals and unpaid volunteers. We investigate the overall quality of the produced translations and compare the translations done at differe...
... middle of paper ...
...e seems to be reasonably well represented.
It will be necessary to confirm the experiments with further translations to have a larger test set. A professional translation will be used as a gold standard to have a more confident reference translation for the automatic evaluations.
It would also be interesting to do similar experiments with other more common or more uncommon language pairs for further comparisons.
References
Winter Mason, Duncan J. Watts. 2009. Financial Incentives and the 'Performance of Crowds'. Proceedings of KDD-HCOMP 2009, Paris, France.
Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: a Method for Automatic Evaluation of Machine Translation. Proceedings of the ACL 2002 conference, Philadelphia, PA.
Jason Pontin. 2007. Artificial Intelligence, With Help From the Humans, The New York Times, March 25th, 2007.
business, people. » Blog Archive » Columbian Exchange." Professional Translation Services | Interpreters | Intercultural Communication & Training. http://www.kwintessential.co.uk/articles/colombia/Columbian-Exchange/5062 (accessed April 3, 2012).
Nearly all Haitian immigrants entering the U.S. are poorly educated, illiterate, and speak only Creole, which is seldom seen in written form. Creole is a “pidgin” language, meaning it is a simplified form of a base language with parts of other languages added. These types of languages were frequently used by sailors, pirates, and other trade people to accommodate the span of communication needs they faced. Haitian Creole is thought to have been derived by combining various native African dialects with the French language of their owners. Very few Haitians (10%) can actually speak French, and one’s ability to do so is seen as an indicator of social class. Because of Haitian views that Creole is the language used by the poor and uneducated, many will claim to be able to speak French and become insulted if it is suggested that they speak Creole. This can pose a problem for the healthcare worker trying to find a way to communicate. Often the only interpreters available to a family are their children who have learned English in schools here. This can create conflict within the family therefore a facility provided interpreter usually produces a better outcome. Written materials are often of no use to the Haitian immigrant.
"For the translator, who stands astride two cultures, possesses two different sensibilities, and assumes a double identity" —Husain Haddawy
In the healthcare setting, it is very important to use medical interpreters. Without interpreters, people who speak different languages would not be able to communicate with healthcare professionals. There are many different ways that a language can be interpreted. A couple of those are actual human interpreters, or electronic interpreters. Both are pretty reliable but an actual human is often looked at as the most reliable interpreter. When you have an actual human interpreting, you don’t have to worry to much on things being translated incorrectly. Some benefits of using electronic interpreters would be the unlimited availability of languages, and being able to get the iPad or laptop as soon as you needed it. When you rely on a person to
Links999. Ethical and moral issues regarding artificial intelligence. Links999.org, 24 Apr. 2011. Web. 25 Apr. 2011. .
Hospitals still use family members to interpret for limited English speaking patients. Hospitals use language line, a company the nurse calls to find interpreter or they use a trained staff member to interpret. Some hospitals continue to use Spanish speaking staff that is bilingual and not trained for translating in the medical field. There is a need for better trained professional interpreters for both the patient and the health care provider. These are necessary components in providing language access in other areas of the united sates with the increasing Spanish-speaking populations (Martinez-Gibson, & Gibson, 2007). Does staff in Emergency Departments continue to use family members and untrained staff as interpreters? Language line and trained interpreters are the only acceptable interpreters (Martinez-Gibson, & Gibson, 2007). When interpreter is needed in an urgent case there needs to be a trained interpreter on staff 24 hours in the emergency department, language line is not always congruent to life saving care that is needed...
WIley, Terrence. "A Languages for Jobs Initiative." Council on Foreign Relations. Council on Foreign Relations, June 2012. Web. 04 May 2014. .
Bilton, Nick. “Artificial Intelligence as a Threat.” The New York Times. The New York Times, 5 Nov. 2014. Web. 17 Nov. 2014l.
Schenk, Eric, and Claude Guittard. "Crowdsourcing: What Can Be Outsourced to the Crowd, and Why?" University of Strasbourg Graduate School of Science and Technology (2009): 1-29. Web.
...slators will not ask for a brief even if they know it is better to have one when they do not receive one. There are two main reasons for this phenomenon. Firstly, considering the tight deadline and wage, translators cannot afford the time to ask for information about the target audience or the communicative purposes and then wait for a couple of days to receive replies from clients. Secondly, translators will not ask for a brief because clients do not know the importance of translation brief and sometimes they will even be annoyed if being asked too many questions. To get more jobs in the future, translators would like to maintain a good relationship with clients. (Jensen, 2009) As we can see, although the Skopos theory stresses the importance of translation brief, the reality in the translation industry does not live up to what is expected in the academic field.
Crevier, D. (1999). AI: The tumultuous history of the search for Artificial Intelligence. Basic Books: New York.
In this essay we are going to study the translation equivalents and the gaps raised from the non-equivalence at word level; then we will analyze some useful strategies for the translation process.
We have build solid reputation in the industry through hard work and innovative products. LinguaSoft has developed a wide range of English learning and test preparation products that help individuals improve their English skills for the real world. Our various products are developed on the principles of “Natural Immersion Program”. We take candidates through a series of practice exercises, fun games and interactive graphics to keep the learning rate of the candidate’s constant throughout the entire learning process. Our various products help candidates develop listening, reading, writing and speaking abilities of the test
The field of Computational Linguistics is relatively new; however, it contains several sub-areas reflecting practical applications in the field. Machine (or Automatic) Translation (MT) is one of the main components of Computational Linguistics (CL). It can be considered as an independent subject because people who work in this domain are not necessarily experts in the other domains of CL. However, what connects them is the fact that all of these subjects use computers as a tool to deal with human language. Therefore, some people call it Natural Language Processing (NLP). This paper tries to highlight MT as an essential sub-area of CL. The types and approaches of MT will be considered, and limitations discussed.
Artificial Intelligence “is the ability of a human-made machine to emulate or simulate human methods for the deductive and inductive acquisition and application of knowledge and reason” (Bock, 182). The early years of artificial intelligence were seen through robots as they exemplified the advances and potential, while today AI has been integrated society through technology. The beginning of the thought of artificial intelligence happened concurrently with the rise of computers and the dotcom boom. For many, the utilization of computers in the world was the most advanced role they could ever see machines taking. However, life has drastically changed from the 1950s. This essay will explore the history of artificial intelligence, discuss the