A Review of Translation Tools From a Post-editing Perspective

Is machine translation post-editing worth the effort? A survey of inquiry into post-editing and effort

Maarit Koponen, University of Helsinki

ABSTRACT

Advances in the field of motorcar translation have recently generated new interest in the employ of this technology in various scenarios. This development raises questions over the roles of humans and machines as machine translation appears to be moving from the peripheries of the translation field closer to the middle. The situation which influences the work of professional translators about involves post-editing motorcar translation, i.e. the use of automobile translations as raw versions to be edited past translators. Such exercise is increasingly commonplace for many language pairs and domains and is likely to course an even larger part of the work of translators in the time to come. While recent studies indicate that post-editing loftier-quality machine translations can indeed increase productivity in terms of translation speed, editing poor machine translations tin be an unproductive task. The question of attempt involved in post-editing therefore remains a central issue. The objective of this article is to present an overview of the use of post-editing of auto translation as an increasingly primal practice in the translation field. Based on a literature review, the commodity presents a view of current knowledge concerning post-editing productivity and effort. Findings related to specific source text features, as well as machine translation errors that appear to be connected with increased post-editing effort, are also discussed.

KEYWORDS

Machine translation, MT quality, MT post-editing, post-editing try, productivity.

1. Introduction

In recent years, the translation industry has seen a growth in the amount of content being translated likewise as pressure to increase the speed and productivity of translation. At the same fourth dimension, technological advances in the field of automobile translation (MT) development take led to wider availability of MT systems for diverse language pairs and improved MT quality. Combined, these factors have contributed to a renewed surge of interest in MT both in enquiry and practice. This changing mural of the translation industry raises questions on the roles of humans and machines in the field. For a long time, MT may have seemed relatively peripheral, with only express employ, but these recent advances are making MT more central to the translation field.

On the ane hand, the availability of complimentary online MT systems is enabling apprentice translators as well as the full general public to use MT. For example, the European Committee Directorate Full general for Translation (DGT) decided in 2010 to develop a new MT organization intended not only every bit a tool for translators working at the DGT, but also for their customers for producing raw translations as self-service (Bonet 2013: 5). On the other hand, MT is too spreading in more professional contexts as it becomes integrated in widely used translation retention systems. While MT fully replacing human being translators seems unlikely, human-machine interaction is certainly becoming a larger part of the piece of work for many professional translators.

Despite improved MT quality in many language pairs, machine-translated texts are more often than not notwithstanding far from publishable quality, except in some limited scenarios involving narrow domains, controlled languages and dedicated MT systems. Therefore, a mutual practice for including MT in the workflow is to utilise motorcar translations as raw versions to be farther postal service-edited by human translators. This growing practical involvement in the use of mail-editing of MT is also reflected on the enquiry side. Specific post-editing tools are being developed and the practices and processes involved are being studied by various researchers. Involvement in post-editing MT is also reflected by a number of recent publications including an edited volume (O'Brien et al. 2014), a special periodical effect (O'Brien and Simard 2014) and international inquiry workshops focusing on the subject field: the Workshops on Mail-editing Technology and Practise that have been organised annually since 2012 are but i instance. Information about the latest workshop in 2015 is available at https://sites.google.com/site/wptp2015/.

Studies take shown that post-editing high-quality MT can, indeed, increase the productivity of professional translators compared to manual translation 'from scratch' (see, for example, Guerberof 2009, Plitt and Masselot 2010). Notwithstanding, editing poor machine translation remains an unproductive job. For this reason, 1 of the central questions in post-editing MT is how much effort is involved in postal service-editing and the amount of effort that is acceptable: when is using MT actually worth information technology? Estimating the actual effort involved in mail-editing is important in terms of the working weather condition of the people involved in these workflows. Equally Thicke (2013: 10) points out, the post-editors are ultimately the ones paying the price for poor MT quality. Objective means of measuring effort are also important in the pricing of post-editing work. Pricing is often based on either hourly rates or determining a scale similar to the fuzzy match scores employed with translation retention systems (Guerberof Arenas 2010: 3). One problem with these practices is that measuring bodily post-editing time is not always simple and the number of changes needed may not be an accurate measure of the endeavor involved.

This paper presents a survey of research investigating the postal service-editing of MT and, in particular, the effort involved. Section 2 provides an overview of the history and more recent developments in the applied use of MT and mail-editing. Section 3 presents a survey of the literature on post-editing in general, as well as a closer look into postal service-editing effort. Section 4 discusses the role of MT and mail-editing every bit they evolve from a peripheral position towards a more than central practise in the linguistic communication industry.

two. Post-editing in exercise – some history and current state

While the surge of interest in MT and post-editing workflows appears recent, it is in fact not a new issue; rather, postal service-editing is ane of the earliest uses envisioned for MT systems. In his historical overview of post-editing and related research, García (2012: 293) notes that "[p]ostediting was a surprisingly hot topic in the belatedly 1950s and early 1960s." Edmundson and Hays (1958) outline 1 of the showtime plans for an MT system and post-editing process, which was intended for translating scientific texts from Russian into English at the RAND Corporation. In this early arroyo, information technology was assumed that the mail-editor would work on the machine-translated text with a grammer lawmaking indicating the lexical category, case, number and other details of every word, and therefore would need "to have skilful command of English grammar and the technical vocabulary of the scientific manufactures existence translated" (Edmundson and Hays 1958: 12). However, command of the source language was not required. According to García (2012: 294), this commencement documented post-editing program was discontinued in the early 1960s. In the 1960s, mail service-editing was used also by the US Air Force's Foreign Engineering Partitioning and Euratom, but funding in the The states concluded in function because the written report past the Automatic Linguistic communication Processing Committee in 1966 (the so-chosen ALPAC written report) institute mail-editing not to be worth the effort in terms of time, quality and difficulty compared to human translation (García 2012: 295-296).

While early uses of both MT as a applied science and post-editing every bit a exercise failed to alive up to the initial visions, evolution connected. MT systems and postal service-editing processes were implemented in organisations such as the European union and the Pan-American Wellness Organization besides as in some companies from the 1970s onwards (García 2012: 296-299). Equally studies accept shown that MT tin can at present produce sufficient quality to be usable in business contexts and increase the productivity of translators at least for some linguistic communication pairs, mail-editing workflows have become increasingly common (run into, for example, Plitt and Masselot 2010, Zhechev 2014, Silva 2014). A recent study past Gaspari et al. (2015) gives an overview of various surveys carried out in recent years, in addition to reporting their own survey of 438 stakeholders in the translation and localisation field. According to Gaspari et al. (2015: thirteen–17), xxx% of their respondents were using MT and 21% considered it sure or at least probable that they would start using it in the future. Of those using MT, 38% answered that it was always post-edited, 30% never performed mail service-editing, whereas the remaining 32% used mail service-editing at varying frequencies. Overall, the surveys cited point to increasing use of MT and post-editing, increased demand for this service and expectations that the trend volition proceed.

As an example of the productivity gains often quoted, Robert (2013: 32) states that post-editing MT tin can increase the boilerplate number of words translated by a professional from 2,000 to three,500 words per mean solar day. According to Guerberof Arenas (2010: 3), a productivity gain of five,000 words per day is unremarkably quoted, but she points out that the actual figures volition vary across different projects and post-editors. García (2011: 228) also notes that this effigy involves certain conditions — professional, experienced post-editors, domain-specific MT systems and source text potentially pre-edited for MT — which may not always apply.

The employ, and usability, of MT and mail service-editing also varies greatly in different linguistic communication pairs. Discussing a survey of translators at the European Committee Advisers Full general for Translation, Leal Fontes (2013) presents the varying uptake rates for various linguistic communication combinations. For case, translators working with the language pairs French-Spanish, French-Italian and French-Portuguese rated nigh MT segments reusable, but for English language-German language, English language-Finnish and English-Estonian, translators considered MT simply sufficient to suggest ideas for expressions or not usable at all (Leal Fontes 2013: eleven). Some languages may nowadays their own challenges: languages with rich morphology, such every bit Finnish, are known to be difficult for MT systems. Furthermore, language pairs with modest markets frequently attract less interest in development piece of work. On the user side, productivity studies at large companies like Autodesk have shown great variation in the gains achieved and corporeality of editing needed (Plitt and Masselot 2010, Zhechev 2014). Most of the applied use of post-editing involves professional contexts, but crowdsourcing has also been explored, for example, in disaster relief situations (Hu et al. 2011), websites (Tatsumi et al. 2012), and discussion forum contexts (Mitchell et al. 2013).

Every bit post-editing workflows become increasingly common, there is a growing need for both organisation-specific and more general post-editing guidelines (see, for case, Guerberof 2009, 2010). A set of such guidelines by TAUS (2010), intended to help customers and linguistic communication service providers in setting articulate expectations and instructing mail service-editors, includes general recommendations for decreasing the amount of mail service-editing needed, equally well as basic guidelines for carrying out post-editing to 2 divers quality levels. The first level of quality, termed 'good enough', is defined equally comprehensible and accurate so that information technology conveys the significant of the source text, without necessarily being grammatically or stylistically perfect. At this level, the postal service-editor should make sure that the translation is semantically correct, that no information is accidentally added or omitted and that it does not contain offensive or inappropriate content. The second level would be 'publishable quality,' like to that expected of human translation. In addition to being comprehensible and accurate, information technology should also exist grammatically and stylistically advisable. An International Standard (ISO/CD 18587:2014) is also being drafted with a view to specifying standardised requirements for the postal service-editing procedure and post-editor competences. This draft standard likewise specifies two post-editing levels (lite and full mail-editing), similar to the two levels of the TAUS guidelines (ISO/CD 18587: vi). An earlier definition with iii levels (rapid, minimal and full post-editing) can exist institute in Allen (2003).

3. Post-editing research

The first report focusing on postal service-editing can be dated to 1967 (García 2012: 301). This study, conducted by Orr and Minor (1967), compared reading comprehension by exam subjects who read a raw machine translation, a mail-edited machine translation or a transmission translation of scientific texts. The test measured the pct of correct answers, fourth dimension taken to terminate the comprehension examination and efficiency, which was measured as the number of correct answers per 10-minute fourth dimension interval. While they found some differences between postal service-edited and manually translated texts, their results indicated that differences were much larger between reading machine-translated and post-edited versions (Orr and Small 1967: ix). For a more comprehensive historical perspective on post-editing studies run into García (2012). More than recently, research both on the bookish and the commercial sides has focused on productivity: mail service-editing speed or number of pages post-edited in a given timeframe compared to translation from scratch, oft combined with some class of quality evaluation. Post-editing effort, which tin can exist divers in different means (see Department iii.4), has also received much interest lately.

The following sub-sections nowadays an overview of post-editing studies. Sub-department iii.1 discusses productivity studies and sub-section three.ii presents studies related to the quality of the post-edited texts. Sub-department three.3 presents studies investigating the viability of post-editing without access to the source text, i.due east. so-chosen monolingual post-editing. Sub-department 3.4 addresses the concept of post-editing effort, ways of measuring it and findings related to features potentially affecting the corporeality of effort.

three.i. Productivity

Extensive productivity studies for post-editing MT compared to manual translation have been carried out in diverse companies. A study at Autodesk, reported by Plitt and Masselot (2010), involved twelve professional translators working from English into French, Italian, German or Spanish and compared their throughput (words translated per hour) when post-editing MT to manual translation. On the boilerplate, mail-editing was plant to increase throughput by 74%, although there was considerable variation between different translators (Plitt and Masselot 2010: x). The continuation study reported by Zhechev (2014) included further target languages and showed large variation in productivity gains betwixt different linguistic communication pairs.

Past dissimilarity, studies by Carl et al. (2011) and García (2010) did non discover significant improvements in productivity when comparing post-editing and transmission translation. Yet, Carl et al. (2011: 137-138) acknowledge that their test subjects did not have experience in post-editing or in the utilise of the tools involved, and the participants in García'southward study (2010: seven) were trainee translators, which may have some connection to the lack of productivity gains. While information involving relatively inexperienced subjects may not entirely reflect the processes and productivity of experienced, professional post-editors, it should be noted that Guerberof Arenas (2014) plant no pregnant differences in processing speed when comparing experienced versus novice post-editors.

In general, studies accept shown considerable variation between editors in the speed of post-editing (Plitt and Masselot 2010, Sousa et al. 2011) and mail-editors appear to differ more in terms of mail service-editing time and number of keystrokes than in terms of number of textual changes made (Tatsumi and Roturier 2010, Koponen et al. 2012). These studies have besides institute varying editor profiles: speed and number of keystrokes or the number of edits made exercise not necessarily correlate and post-editors accept different approaches to the mail service-editing process.

In summary, research into the productivity of post-editing shows partly conflicting results: some studies have reported quite notable productivity gains when comparison post-editing to manual translation, while others have not found significant differences. The productivity gains appear to depend, at least partly, on specific conditions. These weather condition chronicle to sufficiently high-quality motorcar translation, which is currently achievable for sure language pairs, and machine translation systems geared toward the specific text type existence translated. The post-editors' familiarity with the tools and processes involved in the post-editing job may also play a part. In add-on, comparison of the post-editors themselves shows private variation: the differing work processes of different people appear to affect how much they benefit from the use of motorcar translation.

3.2. Quality

Studies have also addressed quality, every bit there have been concerns that the quality of post-edited texts would exist lower than that of manually translated texts. Fiederer and O'Brien (2009) addressed this question in an evaluation experiment that compared a prepare of sentences translated either manually or by post-editing MT. The unlike versions were rated separately for clarity (intelligibility of the sentence), accuracy (aforementioned meaning as the source sentence) and style (appropriateness for purpose, naturalness and idiomaticity). Fiederer and O'Brien (2009: 62-63) plant that mail-edited translations were rated higher for clarity (although only very slightly) and accuracy, only lower in terms of style. When asked to select their favourite translated version of each judgement, the evaluators mostly preferred the manually translated versions, which may signal that they were most concerned with way. Using a slightly unlike arroyo, Carl et al. (2011) had evaluators rank manually translated and post-edited versions of sentences in order of preference, and found a slight, although not significant, preference for the post-edited translations.

Some studies have approached quality in terms of the number of errors found in the translations. In Plitt and Masselot's (2010) study, both the manually translated and post-edited texts were assessed by the company'due south quality balls team according to the criteria applied to all their publications. All translations, whether manually translated or postal service-edited, were deemed acceptable for publication, simply, somewhat surprisingly, the evaluators in fact marked more manually translated sentences as needing corrections (Plitt and Masselot 2010: xiv-15). In another fault-based study, García (2010) used the guidelines by the Australian National Accreditation Authorization for Translators and Interpreters to compare transmission and post-edited translations. Post-edited MT and manually translated texts received comparable evaluations, with mail-edited texts in fact receiving slightly higher average marks from both of the evaluators involved in the written report (García 2010: xiii-15).

Based on the studies assessing the quality of post-edited texts, it appears that mail-editing tin lead to quality levels similar to manually translated texts. In fact, depending on the quality attribute examined, mail service-edited texts are sometimes even evaluated equally better.

3.3. Monolingual post-editing

While post-editing practice and research mainly involve bilingual post-editors correcting the MT based on the source text, some studies have as well investigated scenarios where a post-editor would work on the MT without admission to the source text. In these scenarios, the central question becomes whether monolingual readers are able to understand the source meaning based on a automobile translation lone. Koehn (2010) describes a fix-upwards where examination subjects mail-edited MT without access to the source text. The correctness of the post-edited sentences was and so evaluated based on a unproblematic standard of right or wrong, where a right sentence was defined equally "a fluent translation that contains the same meaning in the certificate context" (Koehn 2010: 541). Depending on the language pair and MT system, betwixt 26% and 35% of the sentences in Koehn's study were accepted as correct. A like approach was utilised past Callison-Burch et al. (2010), whose results show great variation in the percentages of sentences accustomed as correct, depending on the language pair and organisation used. For the best system in each language pair investigated, the rates of sentences evaluated as correct ranged from 54% to fourscore%, whereas for the everyman-rated systems, in some cases less than 10% of the sentences were accepted as correct after postal service-editing (Callison-Burch et al. 2010: 28). One difficulty with the evaluation against this standard is that it is not articulate whether sentences were rejected due to errors in language or in meaning.

Other monolingual postal service-editing studies have utilised split up evaluation for language and significant. In their study of monolingual post-editing of text messages for emergency responders in a catastrophe situation, Hu et al. (2011) carried out an evaluation on carve up five-bespeak scales for fluency of language and adequacy of meaning; 24% to 39% of sentences (depending on system and test fix) achieved the highest adequacy score as rated by two evaluators. For language, the percentage of sentences achieving the highest fluency score ranged from under 10% to over thirty% (Hu et al. 2011: 401-403). A report carried out by Koponen and Salmi (2015) besides investigated the question of linguistic communication and content errors in a monolingual mail-editing experiment. The postal service-edited versions were evaluated separately for right language and correct meaning, and the results show that test subjects were able to correct about 30% of the sentences for both language and meaning, with an additional 20% corrected for meaning merely still containing language errors. The results reported in Koponen and Salmi (2015: 131) suggest that many of the remaining language errors, such every bit punctuation or spelling errors, appear to be due to inattentiveness on the part of at least some test subjects. In some cases, the corrections themselves independent errors ranging from misspellings to grammer issues caused by correcting merely office of the sentence.

The success of monolingual mail service-editing has also been direct compared to bilingual postal service-editing, where the post-editors accept access to the source text. Mitchell et al. (2013) carried out such a comparing in their study of machine-translated online user forum posts and evaluated quality with regard to fluency, comprehensibility and fidelity of significant by assessing whether the post-editing improved or possibly lowered the quality scores. The results showed variation in the unlike quality measures. Monolingual post-editing was establish to improve the allegiance scores less than bilingual post-editing (43% vs 56% of sentences) in the English-German language pair, only slightly more often (67% vs 64% of sentences) in the English-French language pair. For improving comprehensibility, monolingual post-editing was in both cases less successful than bilingual: in the English-German language pair, 57% of sentences improved in monolingual mail-editing compared to 64% of sentences in the bilingual condition, a range that was smaller than in the English-French pair, where monolingual mail-editing was found to meliorate only 48% of sentences, compared to 63% of sentences improved by bilingual post-editing. In terms of fluency, monolingual and bilingual post-editing were equally successful in the English-French pair (63% of sentences improved in both conditions), and very similar in the English language-German pair (67% vs 70% of sentences).

Although early researchers envisioning mail service-editing of MT, like Edmundson and Hays (1958), assumed simply monolingual mail-editing would be necessary, these later studies do non support its feasibility. Even if monolingual mail-editors are able to improve the language, at to the lowest degree and then far MT quality has not been sufficient to reliably convey the meaning of the source text. In many cases, it even seems that the mail-editors are not very attentive to linguistic communication errors either.

iii.4. Post-editing endeavor

While much of the involvement, specially on the manufacture side, focuses on the speed or number of pages produced through post-edited automobile translation compared to manual translation, this aspect reflects only function of the effort involved. In ane of the most comprehensive early studies on post-editing, Krings (2001: 178-182) distinguishes three split but interlinked dimensions of post-editing effort: temporal, technical, and cognitive. The temporal aspect of post-editing is formed by the technical operations (insertions, deletions and reordering of words) necessary for corrections, as well as the cognitive effort involved in detecting the errors and planning the corrections. When attempting to measure the effort involved in post-editing and to determine the features that potentially atomic number 82 to increased effort, scholars have used different approaches.

One possible way of investigating post-editing effort is, naturally, to enquire the people involved to appraise how much try they perceive or expect during the post-editing of a given text. With the intention of creating a specific post-editing attempt evaluation method, Specia et al. (2010: 43) advise a four-signal calibration to be used by professional translators for rating automobile-translated sentences to bespeak the need for post-editing. A slightly different scale was used by Callison-Burch et al. (2012), who asked editors, afterward post-editing, to evaluate on a 5-signal scale how much of the judgement needed to be edited. The results of such evaluations obviously vary greatly depending on the quality of the MT. For example, in the data used by Specia et al. (2010: 43) average scores for iv,000 sentences translated from English to Spanish ranged from 1.3 to ii.8 on a scale from i (total retranslation necessary) to four (no mail-editing needed). scores given by human evaluators are often considered complicated (see, for example, Callison-Burch et al. 2012: 25) because they are to some extent subjective, and different people may have different perceptions of the effort involved. A written report by Sousa et al. (2011) has, however, shown that sentences requiring less time to edit are more than often tagged as low attempt by evaluators. The human scores may therefore exist useful in providing a rough thought of whether the output of a given car translation organisation is suitable for post-editing.

In many cases, the determination of effort is based on post-editing fourth dimension. Although time tin can be considered the most visible part of postal service-editing endeavor (Krings 2001: 179), collecting authentic information almost post-editing time is not necessarily piece of cake in applied work contexts. For example, mail service-editors' self-reports of the time used may not be complete or sufficiently detailed, and recording accurate information would require specialised tools that are not e'er available (for a more detailed discussion, run into Moran et al. 2014). Interest in collecting data for investigating post-editing effort has led to the development of various tools (see, for example, Aziz et al. 2012, Elming et al. 2014, Moran et al. 2014), which include unlike functionalities to record keylogging and sometimes eye tracking data in addition to time.

Technical try can exist measured using computerised metrics, such as the usually used Human-targeted Translation Edit Charge per unit or HTER (Snover et al. 2006). These metrics compare how many words have been changed (added, deleted, replaced or moved) between the auto-translated version and the post-edited version of a given sentence, and thereby reflect the technical endeavour to some extent. Nonetheless, every bit Elming et al. (2014) betoken out, the use of keylogging to track how the changes were made is a better mensurate of the actual technical effort. Furthermore, the number of changes does not always stand for to the fourth dimension needed for post-editing. For example, Tatsumi (2009) institute that mail service-editing time is sometimes longer or shorter than expected based on the number of edits.

Much attention has recently been paid to which factors influence post-editing try. Tatsumi (2009) suggested source sentence length and structure, every bit well as specific types of errors, every bit possible explanations for cases where post-editing takes longer than expected based on the number of edits. Studies accept since explored what specific types of errors in the machine-translated output or specific source text characteristics might be linked to increased effort. Temnikova (2010) examined different types of edits in relation to post-editing time and establish that machine translations of texts that had been simplified using controlled language rules were faster to edit; they independent more edits assumed to be elementary, such as changing the morphological class of a word, and less edits considered to be more cognitively enervating, such as correcting the word guild or editing mistranslated idioms (Temnikova 2010: 3488-3489). Other studies comparing sentences with human evaluation scores or mail-editing times have besides found that sentences involving less effort, equally indicated by higher human scores or shorter post-editing times, tend to contain more than edits related to word forms and unproblematic substitutions of words of the same give-and-take grade, while sentences with low scores or long post-editing times involve more edits related to give-and-take order, edits where the word class was changed, and corrections of mistranslated idioms (Koponen 2012, Koponen et al. 2012).

In studies concerned with the effect of source text features and their relation to mail service-editing attempt, sentence length, in particular, has been constitute to take some effect. Both very long and very brusque sentences have been plant to have longer post-editing times than expected based on the number of edits (Tatsumi 2009, Tatsumi and Roturier 2010), and very long sentences tend to score low in human evaluations (Koponen 2012). Some other gene may be sentence structure. Incomplete sentences as well every bit complex and chemical compound sentences were found to be associated with longer mail service-editing times by Tatsumi and Roturier (2010: 47-49). In a report investigating specific source text patterns within sentences, Aziz et al. (2014) found that at that place appears to be some connection between longer postal service-editing times and edits related to passages containing specific types of verbs or substantive sequences. Vieira (2014), on the other hand, institute some connection betwixt increased cognitive effort and prepositional phrases, as well every bit instances where repetitions of the same words appear within a sentence.

Overall, cognitive effort appears hard to capture. In an early approach, Krings (2001) utilised think-aloud protocols (TAP), where the mail service-editors reported their actions verbally during the post-editing process. While this approach has been widely used in translation process research in the past, information technology has many drawbacks, such as slowing downward the process and changing the cerebral processing involved (for a full discussion, see O'Brien 2005: 41-43). Pauses found in keylogging data recorded during post-editing accept attracted attention equally a measure of cognitive effort. O'Brien (2005) institute some correlation between potentially hard source text features identified past comparison different post-editors' versions and pauses recorded in keyboard logging. In a farther written report, O'Brien (2006) analysed pauses during the mail-editing of sentences with source text characteristics previously shown to involve increased effort and sentences without these features, but institute difficulties in connecting the location and duration of pauses to the cognitive processing. While these studies and much of the earlier research involves identifying long pauses, a unlike arroyo which examines clusters of short pauses has been explored by Lacruz et al. (2012) and Lacruz and Shreve (2014). The rationale, as described by Lacruz and Shreve (2014: 263), is that segments requiring more cognitive mail service-editing attempt would contain a higher density of short pauses than segments requiring little effort. They identify edits related to wrong syntax and mistranslated idioms as some of the instances where increased cognitive effort is required (Lacruz and Shreve 2014: 267).

In summary, the effort involved in post-editing tin be investigated in terms of post-editing fourth dimension, the technical endeavour of performing the corrections and the try perceived past humans. Studies related to these aspects have so far identified some features which appear connected to increased attempt. They may chronicle to source text characteristics, such every bit the length and structure of the sentences, or specific features like noun sequences. Others relate to errors in the motorcar-translated output, leading to edits involving word order and correction of mistranslated idioms, for example. Questions nonetheless remain in determining the amount of endeavor involved in postal service-editing, particularly the amount of cognitive effort. New technology like heart tracking, which is based on the supposition that instances where the gaze fixates on a word for a long duration or multiple times signal increased cerebral attempt (see Carl et al. 2011: 138), has been used in recent studies (e.k. Vieira 2014). It may, in the futurity, offering further information nigh the features affecting effort.

4. MT and post-editing — from the peripheries towards the center

The championship of this paper posed the question of whether MT and post-editing are worth the effort involved. To investigate this question, an overview was presented of studies related to diverse aspects of the employ of MT, post-editing and the effort involved. Based on this review, the answer appears to exist: yes, when used in suitable conditions. The overall pattern arising from the literature seems to exist that post-editing machine translation can increase the productivity (Department 3.1) of translators in terms of speed, while retaining or in some cases fifty-fifty improving the quality of their translations (Section 3.2). Yet, such benefits are not ever guaranteed. Rather, they are dependent on the quality of the machine translations, as can be seen from the widely varying results for different automobile translation systems, different texts and language pairs. In that location as well appear to exist other factors involved. For example, individual translators' work processes and feel may affect whether, and to what extent, they benefit from the machine translation as a raw version. In addition, post-editing does not currently appear viable if the post-editors have no admission to the source text (Section three.iii), although this has been suggested equally a potential scenario. Enquiry has too identified some potential features affecting the amount of attempt required in mail service-editing (Section 3.iv), but questions still remain every bit to how to accurately determine effort.

The overview of the use of auto translation, mail service-editing and research on these problems also shows how postal service-editing motorcar-translated texts is increasingly becoming a part of the translation workflow, simultaneously prompting new research interests. In this manner, MT and postal service-editing are moving from a rather peripheral position to a more central place in the translation field in different ways: as a practice within the translation industry, a inquiry topic for Translation Studies and a topic for translator training. This motion to a more than central position is perhaps nearly evident in . Equally seen in Department 2, postal service-editing of MT is indeed already an established role of the translation workflow in many professional person contexts. While MT quality, and therefore as well the feasibility of mail-editing, still varies in different language pairs, with efforts by organisations such every bit the European Commission likewise equally linguistic communication service providers, the development seems likely to continue. Inquiry into MT and post-editing has for a long time focused mainly on the technical side of MT system evolution, while in the field of Translation Studies, MT has been a rather peripheral issue. The situation has, however, been changing, with, for example, Fiederer and O'Brien (2009: 52) arguing for the need to engage researchers in the Translation Studies community more than actively in the MT field. As human-car interaction has increased in professional practice, post-editing processes and try in detail accept, indeed, recently go more central areas of involvement to translation scholars.

The increased applied use suggests that at least in some contexts machine translation and post-editing certainly has been found to be worthwhile. At the same fourth dimension, research has too provided show that mail-editing machine-translated output can increase productivity in situations suited to the post-editing process and in this style be worth the attempt. However, post-editing does not appear to be suited for all scenarios. One of the primal questions that remains is how to determine when post-editing is, or is non, worth the attempt. Authentic measurement of the bodily endeavour would be important as it has implications not only for productivity, simply besides for the working atmospheric condition of the mail-editors.

The increased use of machine translation and post-editing workflows has already changed the part of humans and machines in the field, and will likely go on to do then.

Bibliography
  • Allen, Jeff (2003). "Post-editing." Harold Somers (ed.) (2003) Computers and Translation. A Translator's Guide. Amsterdam/Philadelphia: John Benjamins, 297–317.
  • Aziz, Wilker, Sheila C. M. Sousa and Lucia Specia (2012). "PET: A Tool for Post-editing and Assessing Machine Translation." Nicoletta Calzolari et al. (eds) (2012) Proceedings of the eighth International Briefing on Linguistic communication Resource and Evaluation. European Language Resources Association (ELRA),3982–3987.
  • Aziz, Wilker, Maarit Koponen and Lucia Specia (2014). "Sub-judgement Level Analysis of Motorcar Translation Post-editing Attempt." O'Brien et al. (2014), 170–199.
  • Bonet, Josep (2013). "No rage confronting the auto." Languages and Translation 6, four–5. http://ec.europa.eu/dgs/translation/publications/magazines/languagestranslation/
    documents/issue_06_en.pdf (consulted xi.6.2015).
  • Callison-Burch, Chris, Philipp Koehn, Christof Monz, Kay Peterson, Mark Przybocki and Omar F. Zaidan (2010). "Findings of the 2010 joint workshop on statistical auto translation and metrics for machine translation." Chris Callison-Burch et al. (eds) (2010) ACL 2010: Joint Fifth Workshop on Statistical Motorcar Translation and MetricsMATR. Proceedings of the workshop. Stroudsburg, PA: Association for Computational Linguistics, 17–53.
  • Callison-Burch, Chris, Philipp Koehn, Christof Monz, Matt Post, Radu Soricut and Lucia Specia (2012). Chris Callison-Burch et al. (eds) (2012) "Findings of the 2012 workshop on statistical machine translation." 7th Workshop on Statistical Car Translation. Proceedings of the workshop. Stroudsburg, PA: Association for Computational Linguistics, 10–51.
  • Carl, Michael, Barbara Dragsted, Jakob Elming, Daniel Hardt and Arnt Lykke Jakobsen (2011). "The procedure of post-editing: A airplane pilot study." Bernadette Sharp et al. (eds) (2011) Proceedings of the 8th International NLPSC workshop. Special theme: Human-machine interaction in translation. Copenhagen Studies in Language 41. Frederiksberg: Samfundslitteratur, 131–142.
  • Edmundson, Harold P. and David Thousand. Hays (1958). "Inquiry methodology for machine translation." Mechanical Translation 5(i), viii-15.
  • Elming, Jakob, Laura Winther Balling and Michael Carl (2014). "Investigating User Behaviour in Post-Editing and Translation using the CASMACAT Workbench." O'Brien et al. (2014), 147–169.
  • Fiederer, Rebecca and Sharon O'Brien (2009). "Quality and car translation: A realistic objective?" The Journal of Specialised Translation 11, 52–74. http://www.jostrans.org/issue11/art_fiederer_obrien.pdf (consulted 11.half-dozen.2015).
  • García, Ignacio (2010). "Is auto translation prepare yet?" Target 22(1), vii–21.
  • (2011). "Translating by mail-editing: Is it the mode forward?" Automobile Translation 25(3), 217-237.
  • — (2012). "A brief history of postediting and of research on postediting." Anthony Pym and Alexandra Assis Rosa (eds) (2012) New Directions in Translation Studies. Special Result of Anglo Saxonica 3(3), 292–310.
  • Gaspari, Federico, Hala Almaghout and Stephen Doherty (2015). "A survey of auto translation competences: insights for translation engineering educators and practitioners." Perspectives: Studies in Translatology, 1–26. http://dx.doi.org/10.1080/0907676X.2014.979842 (consulted xi.half dozen.2015).
  • Guerberof, Ana (2009). "Productivity and quality in MT mail service-editing." Marie-Josée Goulet et al. (eds) (2009) Across Translation Memories Workshop. MT Summit XII, Ottawa. Association for Machine Translation in the Americas.
  • Guerberof Arenas, Ana (2010). "Project management and machine translation." Multilingual 21(three), 1–4.
  • — (2014). "The role of professional person experience in post-editing from a quality and productivity perspective." O'Brien et al. (2014), 51–76.
  • Hu, Chang, Philip Resnik, Yakov Kronrod, Vladimir Eidelman, Olivia Buzek and Benjamin B. Bederson (2011). "The value of monolingual crowdsourcing in a real-world translation scenario: simulation using Haitian Creole emergency SMS messages." Chris Callison-Burch et al. (eds) (2011) Proceedings of the 6th Workshop on Statistical Machine Translation (WMT '11). Stroudsburg PA: Association for Computational Linguistics, 399–404.
  • ISO/CD 18587:2014 (2014). Translation Services – Post-editing of automobile translation output – Requirements. International Arrangement for Standardization.
  • Koehn, Philipp (2010). "Enabling monolingual translators: Postal service-editing vs. options." Ron Kaplan et al. (eds) (2010) Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Clan for Computational Linguistics. Stroudsburg PA: Association for Computational Linguistics, 537–545.
  • Koponen, Maarit (2012). "Comparing Human Perceptions of Post-Editing Effort with Post-Editing Operations." Chris Callison-Burch et al. (eds) (2012) seventh Workshop on Statistical Machine Translation. Proceedings of the workshop. Stroudsburg, PA: Clan for Computational Linguistics,181–190.
  • Koponen, Maarit, Wilker Aziz, Luciana Ramos and Lucia Specia (2012). "Post-Editing Time as a Measure of Cerebral Endeavor." Sharon O'Brien, Michel Simard and Lucia Specia (eds) (2012) Proceedings of the AMTA 2012 Workshop on Post-editing Engineering science and Exercise. Association for Auto Translation in the Americas,11–20.
  • Koponen, Maarit and Leena Salmi (2015). "On the correctness of machine translation: A machine translation post-editing task". The Periodical of Specialised Translation 23, 118–136. http://www.jostrans.org/issue23/art_koponen.pdf (consulted 11.6.2015).
  • Krings, Hans P. (2001). Repairing Texts: Empirical Investigations of Motorcar Translation Mail service-Editing Processes (ed. Geoffrey S. Koby). Kent, Ohio: Kent State Academy Press.
  • Lacruz, Isabel, Gregory M. Shreve and Erik Angelone (2012). "Average Interruption Ratio as an Indicator of Cognitive Effort in Mail-Editing: A Case Study." Sharon O'Brien, Michel Simard and Lucia Specia (eds) (2012) Proceedings of the AMTA 2012 Workshop on Post-editing Applied science and Practice. Association for Machine Translation in the Americas,21–30.
  • Lacruz, Isabel and Gregory G. Shreve (2014). "Pauses and Cognitive Effort in Postal service-editing." O'Brien et al. (2014), 246–272.
  • Leal Fontes, Hilário (2013). "Evaluating Machine Translation: preliminary findings from the get-go DGT-broad translators' survey." Languages and Translation 6, 10–xi. http://ec.europa.eu/dgs/translation/publications/magazines/languagestranslation/
    documents/issue_06_en.pdf (consulted iv.5.2015).
  • Mitchell, Linda, Johann Roturier and Sharon O'Brien (2013). "Community-based post-editing of machine-translated content: monolingual vs. bilingual." Sharon O'Brien, Michel Simard and Lucia Specia (eds) (2013) Workshop Proceedings: Workshop on Mail service-editing Technology and Practice (WPTP-ii). Allschwil: The European Association for Motorcar Translation, 35–44.
  • Moran, John, David Lewis and Christian Saam (2014). "Assay of Mail service-editing Data: A Productivity Field Exam Using an Instrumented CAT Tool." O'Brien et al. (2014), 126–146.
  • O'Brien, Sharon (2005). "Methodologies for Measuring the Correlations between Post-Editing Endeavor and Automobile Translatability." Auto Translation 19(i), 37–58.
  • — (2006). "Pauses as Indicators of Cerebral Endeavour in Post-editing Automobile Translation Output." Beyond Languages and Cultures 7, 1–21.
  • O'Brien, Sharon, Laura Winther Balling, Michael Carl, Michel Simard and Lucia Specia (eds) (2014). Post-editing of Auto Translation: Processes and Applications. Newcastle: Cambridge Scholars Publishing.
  • O'Brien, Sharon and Michel Simard (2014). "Introduction to special issue on mail-editing." Machine Translation 28(iii), 159–164.
  • Orr, David B. and Victor H. Modest (1967). "Comprehensibility of Automobile-aided Translations of Russian Scientific Documents." Mechanical Translation and Computational Linguistics 10(i-2), i–x.
  • Plitt , Mirko and François Masselot (2010). "A Productivity Test of Statistical Machine Translation Post-Editing in a Typical Localisation Context." The Prague Bulletin of Mathematical Linguistics 93, vii–16. http://ufal.mff.cuni.cz/pbml/93/fine art-plitt-masselot.pdf (consulted eleven.vi.2015).
  • Robert, Anne-Marie (2013). "Vous avez dit post-éditrice ? Quelques éléments d'un parcours personnel." The Journal of Specialised Translation 19, 29–40. http://www.jostrans.org/issue19/art_robert.pdf (consulted 11.6.2015).
  • Silva, Roberto (2014). "Integrating Post-editing MT in a Professional Translation Workflow." O'Brien et al. (2014), 24–50.
  • Snover, Matthew, Bonnie Dorr, Richard Schwartz, Linnea Micciulla and John Makhoul (2006). "A Written report of Translation Edit Rate with Targeted Human Annotation." Laurie Gerber et al. (eds) (2006) Proceedings of the 7th Briefing of the Clan for Machine Translation in the Americas. Association for Machine Translation in the Americas, 223–231.
  • Sousa, Sheila C. M., Wilker Aziz and Lucia Specia (2011). "Assessing the Post-Editing Effort for Automated and Semi-Automated Translations of DVD subtitles." Galia Angelova et al. (eds) (2011) Proceedings of the Contempo Advances in Natural Linguistic communication Processing Briefing. Hissar, Republic of bulgaria: RANLP 2011 Organising Committee,97–103.
  • Specia, Lucia, Dhwaj Raj and Marco Turchi (2010). "Machine Translation Evaluation Versus Quality Estimation." Automobile Translation 24(1), 39–50.
  • Tatsumi, Midori (2009). "Correlation Betwixt Automatic Evaluation Metric Scores, Postal service-Editing Speed, and Another Factors." Laurie Gerber et al. (eds) (2009) Proceedings of the MT Summit XI. Association for Machine Translation in the Americas,332–339.
  • Tatsumi, Midori and Johann Roturier (2010). "Source Text Characteristics and Technical and Temporal Post-Editing Attempt: What is Their Relationship?" Ventsislav Zhechev (ed.) (2010) Proceedings of the 2d Joint EM+/CNGL Workshop "Bringing MT to the User: Research on Integrating MT in the Translation Manufacture" (JEC 'ten). Association for Machine Translation in the Americas, 43–51.
  • Tatsumi, Midori, Takako Aikawa, Kentaro Yamamoto and Hitoshi Isahara (2012). "How Good Is Crowd Post-Editing? Its Potential and Limitations." Sharon O'Brien, Michel Simard and Lucia Specia (eds) (2012) Proceedings of the AMTA 2012 Workshop on Postal service-editing Technology and Exercise. Association for Machine Translation in the Americas,69–77.
  • TAUS (2010). Machine Translation Post-editing Guidelines. https://www.taus.cyberspace/think-tank/all-time-practices/postedit-best-practices/car-translation-mail-editing-guidelines (consulted xi.vi.2015)
  • Temnikova, Irina (2010). "A Cognitive Evaluation Arroyo for a Controlled Language Post-Editing Experiment." Nicoletta Calzolari et al. (eds) (2010) Proceedings of the seventh International Conference on Language Resources and Evaluation. European Language Resource Association (ELRA), 3485–3490.
  • Thicke, Lori (2013). "The industrial procedure for quality car translation." The Journal of Specialised Translation xix, 8–eighteen. http://world wide web.jostrans.org/issue19/art_thicke.pdf (consulted eleven.6.2015)
  • Vieira, Lucas Nunes (2014). "Indices of cognitive endeavor in automobile translation post-editing." Machine Translation 28(3), 187–216.
  • Zhechev, Ventsislav (2014). "Analysing the Post-Editing of Motorcar Translation at Autodesk." O'Brien et al. (2014), 2–13.
Biography

Koponen portrait

Maarit Koponen has an MA in English Philology (2005) from the Academy of Helsinki. She is currently writing her PhD in Linguistic communication Technology (focusing on machine translation evaluation and post-editing) at the Academy of Helsinki in the Department of Modern Languages. She has taught computer-assisted translation and post-editing at the University of Helsinki and professional translation at the University of Eastern Finland. She has also worked as a professional translator for over five years.
Email: maarit.koponen@helsinki.fi

buchananthersom2002.blogspot.com

Source: https://www.jostrans.org/issue25/art_koponen.php

0 Response to "A Review of Translation Tools From a Post-editing Perspective"

Publicar un comentario

Iklan Atas Artikel

Iklan Tengah Artikel 1

Iklan Tengah Artikel 2

Iklan Bawah Artikel