Next: ZHANG Ruibo Up: EMSAN paper session: Ruibo Previous: EMSAN paper session: Ruibo Contents
Terminology translation as the key element in my research, ``Assisting the Development of the Field of Electroacoustic Music Studies in China'', has been upgraded into terminology research by introducing a new method called CHEARS (China ElectroAcoustic Resource Survey) Dictionary and Machine Translation supported by a China Scholarship Council grant (CSC). The CHEARS Dictionary not only collects terminology from the well-known ElectroAcoustic Resource Site (EARS) and The Pedagogical ElectroAcoustic Resource Site (EARS 2), but also collects the important words from the EARS glossary and EARS 2 tutorials. The terms in the CHEARS dictionary will be given only one, at most two definitions (mostly translations — EARS can have several). Technically, Machine Translation allows for EARS terminology to be separated into single words, by way of searches within the CHEARS Dictionary. If there is an entry of the word in the dictionary, it will be substituted by the counterpart in Chinese; otherwise it will be kept in the original form (English). Of course, grammar is not taken into consideration at this point; the original sentence order in English will be kept prior to human editing.
The two EARS sites are both run by the Music, Technology and Innovation Research Centre, MTI,
at De Montfort University, Leicester, UK; however, EARS 2 takes a different path from the
original EARS initiative which is a multilingual internationally peer-reviewed professional
(multilingual) glossary with an accompanying subject index for electroacoustic (EA) music as
well as a bibliographic resource. The goal of EARS 2 is to assist young people in learning about
and subsequently gaining interest in composing music with sound. The archetype of CHEARS
Dictionary and Machine Translation (MT) also involves a pair of dynamic websites,
CHEARS.info (simplified Chinese and English) and CHEARSdotinfo.co.uk (traditional Chinese
and English), as its method of representation. One is hosted in Beijing and optimised for Chinese
users with a registered ICP license (Internet Content Provider that is a permit issued by the
Chinese Ministry of Industry and Information Technology to permit China-based websites to
operate in China); the other is hosted in London and optimised for European users (and those who
are stronger in simplified Chinese). The latest research environment uses a MySQL database as an
intensive search engine for speeding up the potential latency of Machine Translation. Since the
end of 2014 this system is fully functional; it is not only including a multi-user communication
area, a text resource, events and music criticism sections for presentation, but also turns the
database into a research-oriented one. In other words, the CHEARS Database is triggered online
by researchers through the research-based dynamic website portal (chears.info and
chearsdotinfo.co.uk) in real-time. It has turned a translation project with the presentation-oriented
website into a research-oriented database project presented through a research-based dynamic
website portal. Finally, the processing of database (Microsoft Access and MySQL) and website
built under the C
and .NET Framework is the research itself.
CHEARS Dictionary and MT are mainly applied to the translation of EARS site from English to Chinese as the first goal. There are more than 1,200 words as vocabulary in the CHEARS dictionary at the moment and the number will be increased continuously as the database expands. Of course, it will never be a substitute for human. Therefore, human editing will be applied prior to publication on the sites. It will save a considerable amount of time for translators, proof readers and project consultants base on the original work flow. This new combined work flow (MT plus human editing) can ensure that every term is translated under peer review, and assist researchers not to make any arbitrary decisions regarding terminology. All in all, the CHEARS Dictionary and MT are only suitable for researchers and musicologists who are bilingual in English and Chinese. The content is generated by and for researcher sin EA music precisely and the intention is that the results are totally different from commercial multilingual applications or web-based plug-ins. The whole process is described as MT plus human editing, which is also known as Computer-assisted or Computer-aided Translation (CAT).
In the combination of EARS and EARS2, there are approximately 70,000 words in the CHEARS Database including the description of more than 700 e-a terms. Until now, 27,000 Chinese words have been translated and the description of more than 200 terminologies are depend on traditional method in natural language translated by 26 human translators and 6 proofreaders covering one third of the total data amount. In the CHEARS Dictionary involving MT, it is not sufficient any more to only get 750 terminologies indexed in macro-level; instead it takes any useful words among 70,000 indexed in micro-level as the object of the terminology research. The ratio of 750 terminologies among 70,000 words is incredibly low; this is the most important reason to the difficulty of understanding to the translated material by Chinese readers. This systematic problem exists obviously after having been dealing with the translation outcomes from previous years. Since the grammar has to be taken in account in the process of human translation in natural language, that will get the original context more or less encoded again and again to confuse readers. Instead of being confused by language phenomenon unpredictably, this method of terminology research will solve the systematic problem. The CHEARS Dictionary with MT takes a much closer look at each single word in the atomic level with highest resolution. This talk will include examples demonstrating the method presented in this talk.
adrian 2015-06-03