The Time Constraint in Conference Interpreting: Simultaneous vs. Consecutive

Gumul, Ewa; Łyda, Andrzej

doi:10.2478/v10015-007-0007-1

Cited by 7 publications

(4 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…EVS measures the average time from when the speaker finishes conveying a piece of information to when the audience hears the corresponding translation, which is similar to AL. The typical EVS of professional human interpreters usually ranges from 3 to 6 seconds [26] to achieve high-quality translation.…”

Section: Latencymentioning

confidence: 99%

Impact of Encoding and Segmentation Strategies on End-to-End Simultaneous Speech Translation

Nguyen¹,

Estève²,

Besacier³

2021

Interspeech 2021

View full text Add to dashboard Cite

In this paper, we present Cross Language Agent -Simultaneous Interpretation, CLASI, a high-quality and human-like Simultaneous Speech Translation (SiST) 1 System. Inspired by professional human interpreters, we utilize a novel data-driven read-write strategy to balance the translation quality and latency. To address the challenge of translating in-domain terminologies, CLASI employs a multi-modal retrieving module to obtain relevant information to augment the translation. Supported by LLMs, our approach can generate error-tolerated translation by considering the input audio, historical context, and retrieved information. Experimental results show that our system outperforms other systems by significant margins. Aligned with professional human interpreters, we evaluate CLASI with a better human evaluation metric, valid information proportion (VIP), which measures the amount of information that can be successfully conveyed to the listeners. In the real-world scenarios, where the speeches are often disfluent, informal, and unclear, CLASI achieves VIP of 81.3% and 78.0% for Chinese-to-English and English-to-Chinese translation directions, respectively. In contrast, state-of-the-art commercial or open-source systems only achieve 35.4% and 41.6%. On the extremely hard dataset, where other systems achieve under 13% VIP, CLASI can still achieve 70% VIP. Demonstrations and human-annotated test sets are available at https://byteresearchcla.github.io/clasi. 0 20 40 60 Figure 1: Performance evaluation. CLASI significantly outperforms the leading commercial and opensource systems using a more reliable VIP metric, achieving human interpreter parity.

show abstract

Section: Latencymentioning

confidence: 99%

Impact of Encoding and Segmentation Strategies on End-to-End Simultaneous Speech Translation

Nguyen¹,

Estève²,

Besacier³

2021

Interspeech 2021

View full text Add to dashboard Cite

show abstract

“…A. Gillies 2007: 24, D. Gile 1995/2009, E. Gomul/ A. Łyda 2007 i dlatego droga do osiągnięcia biegłości w jego wykonywaniu jest długa i najczęściej żmudna.…”

Section: Wnioskiunclassified

Tłumaczenie konsekutywne w wykonaniu zaawansowanego studenta kursu t ł umaczenia ustnego – analiza wyników i głównych problemów

Nasiadka¹

2016

View full text Add to dashboard Cite

Niniejsza publikacja jest dostępna na licencji Creative Commons. Uznanie autorstwa-Użycie niekomercyjne-Bez utworów zależnych 3.0 Polska. Pewne prawa zastrzeżone na rzecz autora. Zezwala się na wykorzystanie publikacji zgodnie z licencją -pod warunkiem zachowania niniejszej informacji licencyjnej oraz wskazania autora jako właściciela praw do tekstu. Treść licencji jest dostępna na stronie: http://creativecommons.org/licenses/by-nc-nd/3.0/pl/ Tłumaczenie konsekutywne w wykonaniu zaawansowanego studenta kursu tłumaczenia ustnego -analiza wyników i głównych problemów Abstract: Consecutive interpreting by an advanced interpreting course student: an analysis of the trainee's performance and the main problems In the article, I concentrate on the process of training future consecutive interpreters. In doing this, I refer to my own experience as an interpreter and a university teacher of consecutive interpreting in order to identify some of the main challenges and obstacles that have to be overcome by both the student and his teacher if they want to form a professional being able to effectively render original thoughts and meanings into another language. I touch on the things that negatively affect the performance of the trainees, with stress among them. Finally, I describe my methods that, in my opinion, help the students interpret better.

show abstract

“…Recent tendencies, however, emphasize the work of the interpreter as being grounded in a sociocultural context; it is claimed that much of what is done can only be understood within that context, rather than on cognitive modeling alone. More specifically, researchers working on context-dependent aspects of interpreting have sought to base their insights on data such as appeals to the political and cultural importance of the political settings (Cronin 2002), motivated shifts of footing (Diriker 2004), judicious omissions as ethically enhancing coherence (Viaggio 2002;Gumul and Lyda 2007), and the now numerous surveys of user expectations (for example, Kurz 1993, Moser 1996, Pöchhacker 2002.…”

Section: Cognition Vs Context In Interpreting Studiesmentioning

confidence: 99%

On omission in simultaneous interpreting: Risk analysis of a hidden effort

Pym

2008

Benjamins Translation Library

View full text Add to dashboard Cite

One of the long-standing debates in studies on simultaneous interpreting would pit "contextualists", who see interpreters' performances as being conditioned by contextual determinants, against "cognitivists", who analyze performances in terms of cognitive constraints that would be the same for all professionals, regardless of context. Gile's Effort Models would seem to be very much in the cognitive camp. However, modeling of the resources used when interpreters make omissions suggests that cognitive management may actively respond to contextual factors such as the aims of the discourse, the strategies of the speakers, and the variable risks of the text items. Analysis of the data from one of Gile's experiments indicates that the cognitive management of omissions is indeed highly variable. Omissions that are low-risk for the aims of the discourse occur in a constant background mode, almost without source-text stimuli, such that they are found in repeat performances with similar frequency but in different places. On the other hand, omissions that incur high levels of risk tend to be repaired in repeat performance. This suggests that simultaneous interpreters strive for non-omission only in the case of high-risk contextualization. Further, since their management skills must incorporate enough contextualization for the necessary risk analysis to take place, the cognitive strategies of interpreters should be modeled in the same terms as those of all other linguistic mediators.

show abstract

The Time Constraint in Conference Interpreting: Simultaneous vs. Consecutive

Cited by 7 publications

References 24 publications

Impact of Encoding and Segmentation Strategies on End-to-End Simultaneous Speech Translation

Impact of Encoding and Segmentation Strategies on End-to-End Simultaneous Speech Translation

Tłumaczenie konsekutywne w wykonaniu zaawansowanego studenta kursu t ł umaczenia ustnego – analiza wyników i głównych problemów

On omission in simultaneous interpreting: Risk analysis of a hidden effort

Contact Info

Product

Resources

About