The Changing Face of Conference Interpreting

Citation

Zhang X (2015) The Changing Face of Conference Interpreting. In: Pastor G, Domínguez M, Florido R & Mendaña M (eds.) New Horizons in Translation and Interpreting Studies. The 7th International Conference of the Iberian Association of Translation and Interpreting Studies (AIETI) 2015, Malaga, Spain, 29.01.2015-31.01.2015. Geneva: Editions Tradulex, pp. 255-263. http://www.tradulex.com/varia/AIETI7-proc.pdf

Abstract
Interpreting service is extraordinary expensive but the requirements for interpreting is increasingly huge. For more than 2,500 years history of interpreting, technological innovation is rarely involved in it. It's time to change face for the interpreting. Microsoft's automatic and simultaneous speech translation system and a prototype of lecture simultaneous translation system have been demostrated in the past years. Nara Institute of Science and Technology in Japan developed a speech translation system based on simultaneous interpretation data. However, limited to the unreliable quality of machine translation (MT) output and the stiff style of MT language, in practice, this kind of automatic interpreting is far from trivial to implement. We intend to develop a computer-aided conference interpreting system (CACIS) to integrate video, audio and text translation technologies into a platform to help the interpreters in the booths to complete their interpreting work. CACIS has different focuses in different conference interpreting scenarios. To a simultaneous interpreting (SI), a reliable and real-time automatic sound recognition (ASR) output of the speaker will be a good reference. ASR transcription may help the interpreter on mike to recall the missing keywords. MT technology will provide referential translation of the ASR transcription for the interpreter in real-time. The key points of this stage are to train a reliable ASR system and develop a satisfied MT system. To a consecutive interpreting (CI), CACIS can extract the keywords such as who, when, where, what and how on the event the speaker talking about from the ASR transcription and generate a note for the interpreter automatically. Note-taking will not be a puzzle to the interpreter and he/she will be set free to concentrate on listening without any boring of the balance between note-taking and his/her short memory. The video processing may capture and interpret some speaker's non-linguistic information, such as gesture and emotion recognition. The ASR transcription and MT suggestion provided by SI processing are also be useful in CI stage. To the scenario of interpreting without the limitation of time and place, such as audio/video translation (AVT), subtitling and shadowing, the outputs from SI or CI can be revised carefully in an interactive platform in the mode of post-editing (PE). The post-editor will revise the transcription of ASR, the MT output, generated notes, and non-linguistic transcriptions in an interactive and collaborative manner and on a web-based platform. A crowd sourcing mode is accepted in this stage. With the help of CACIS, conference interpreting is expected to be more informative and acceptable for the audience.

Keywords
Conference interpreting; computer-aided conference interpreting system (CACIS); translation technologies; technical innovation; machine translation