Category: Learning Zone

NEURAL MACHINE TRANSLATION: THE RISING STAR

NEURAL MACHINE TRANSLATION: THE RISING STAR

These days, language industry professionals simply can’t escape hearing about neural machine translation (NMT). However, there still isn’t enough information about the practical facts of NMT for translation buyers, language service providers, and translators. People often ask: is NMT intended for me? How will it change my life?

A Short History and Comparison

At the beginning of time – around the 1970s – the story began with rule-based machine translation (RBMT) solutions. The idea was to create grammatical rule sets for source and target languages, where machine translation is a kind of conversion process between the languages based on these rule sets. This concept works well with generic content, but adding new content, new language pairs, and maintaining the rule set is very time-consuming and expensive.

This problem was solved with statistical machine translation (SMT) around the late ‘80s and early ‘90s. SMT systems create statistical models by analyzing aligned source-target language data (training set) and use them to generate the translation. The advantage of SMT is the automatic learning process and the relatively easy adaptation by simply changing or extending the training set. The limitation of SMT is the training set itself: to create a usable engine, a large database of source-target segments is required. Additionally, SMT is not language independent in the sense that it is highly sensitive to the language combination and has a very hard time dealing with grammatically rich languages.

This is where neural machine translation (NMT) begins to shine: it can look at the sentence as a whole and can create associations between the phrases over an even longer distance within the sentence. The result is a convincing fluency and an improved grammatical correctness compared to SMT.

Statistical MT vs Neural MT

Both SMT and NMT are working on a statistical base and are using source-target language segment pairs as a basis. What’s the difference? What we typically call SMT is actually Phrase Based Statistical Machine Translation (PBSMT), meaning SMT is splitting the source segments into phrases. During the training process, SMT creates a translation model and a language model. The translation model stores the different translations of the phrases and the language model stores the probability of the sequence of phrases on the target side. During the translation phase, the decoder chooses the translation that gives the best result based on these two models. On a phrase or expression level, SMT (or PBSMT) is performing well, but language fluency and grammar is not good.

‘Buch’ is aligned with ‘book’ twice and only once with ‘the’ and ‘a’ – the winner is the ‘Buch’-’book’ combination

Neural Machine Translation, on the other hand, is using neural network-based, deep, machine learning technology. Words or even word chunks are transformed into “word vectors”. This means that ‘dog’ is not only representing the characters d, o and g, but it can contain contextual information from the training data. During the training phase, the NMT system tries to set the parameter weights of the neural network based on the reference values (source-target translation). Words appearing in similar context will get similar word vectors. The result is a neural network which can process source segments and transfer them into target segments. During translation, NMT is looking for a complete sentence, not just chunks (phrases). Thanks to the neural approach, it is not translating words, it’s transferring information and context. This is why fluency is much better than in SMT, but terminology accuracy is sometimes not perfect.

Similar words are closer to each other in a vector space

The Hardware

A popular GPU: NVIDIA Tesla

One big difference between SMT and NMT systems is that NMT requires Graphics Processing Units (GPUs), which were originally designed to help computers process graphics. These GPUs can calculate astonishingly fast – the latest cards have about 3,500 cores which can process data simultaneously. In fact, there is a small ongoing hardware revolution and GPU-based computers are the foundation for almost all deep learning and machine learning solutions. One of the great perks of this revolution is that nowadays, NMT is not only available for large enterprises, but also for small and medium-sized companies as well.

The Software

The main element, or ‘kernel’, of any NMT solution is the so-called NMT toolkit. There are a couple of NMT toolkits available, such as Nematus or openNMT, but the landscape is changing fast and more companies and universities are now developing their own toolkits. Since many of these toolkits are open-source solutions and hardware resources have become more affordable, the industry is experiencing an accelerating speed in toolkit R&D and NMT-related solutions.

On the other hand, as important as toolkits are, they are only one small part of a complex system, which contains frontend, backend, pre-processing and post-processing elements, parsers, filters, converters, and so on. These are all factors for anyone to consider before jumping into the development of an individual system. However, it is worth noting that the success of MT is highly community-driven and would not be where it is today without the open source community.

Corpora

A famous bilingual corpus: the Rosetta Stone

And here comes one of the most curious questions: what are the requirements of creating a well-performing NMT engine? Are there different rules compared to SMT systems? There are so many misunderstandings floating around on this topic that I think it’s a perfect opportunity to go into the details a little bit.

The main rules are nearly the same both for SMT and NMT systems. The differences are mainly that an NMT system is less sensitive and performs better in the same circumstances. As I have explained in an earlier blog post about SMT engine quality, the quality of an engine should always be measured in relation to the particular translation project for which you would like to use it.

These are the factors which will eventually influence the performance of an NMT engine:

Volume

Regardless of you may have heard, volume is still very important for NMT engines just like in the SMT world. There is no explicit rule on entry volumes but what we can safely say is that the bare minimum is about 100,000 segment pairs. There are Globalese users who are successfully using engines created based on 150,000 segments, but to be honest, this is more of an exception and requires special circumstances (like the right language combination, see below). The optimum volume starts around 500,000 segment pairs (2 million words).

Quality

The quality of the training set plays an important role (garbage in, garbage out). Don’t add unqualified content to your engine just to increase the overall size of the training set.

Relevance

Applying the right engine to the right project is the first key to success. An engine trained on automotive content will perform well on car manual translation but will give back disappointing results when you try to use it for web content for the food industry.

This raises the question of whether the content (TMs) should be mixed. If you have enough domain-specific content you shouldn’t necessarily add more out-of-domain data to your engine, but if you have an insufficient volume of domain-specific data then adding generic content (e.g. from public sources) may help improve the quality. We always encourage our Globalese users to try different engine combinations with different training sets.

Content type

Content generated by possible non-native speaking users on a chat forum or marketing material requiring transcreation is always a challenge to any MT system. On the other hand, technical documentation with controlled language is a very good candidate for NMT.

Language combination

Unfortunately, language combination still has an impact on quality. The good news is that NMT has now opened up the option of using machine translation for languages like Japanese, Turkish, or Hungarian –  languages which had nearly been excluded from the machine translation club because of poor results provided by SMT. NMT has also helped solve the problem of long distance dependencies for German and the translation output is much smoother for almost all languages. But English combined with Latin languages still provides better results than, for example, English combined with Russian when using similar volumes and training set quality.

Expectations for the future

Neural Machine Translation is a big step ahead in quality, but it still isn’t magic. Nobody should expect that NMT will replace human translators anytime soon. What you CAN expect is that NMT can be a powerful productivity tool in the translation process and open new service options both for translation buyers and language service providers (see post-editing experience).

Training and Translation Time

When we started developing Globalese NMT, one of the most surprising experiences for us was that the training time was far shorter than we had previously anticipated. This is due to the amazingly fast evolution of hardware and software. With Globalese, we currently have an average training time of 50,000 segments per hour – this means that an average engine with 1 million segments can be trained within one day. The situation is even better when looking at translation times: with Globalese, we currently have an average translation time between 100 and 400 segments per minute, depending on the corpus size, segment length in the translation and training content.

Neural MT Post-editing Experience

One of the great changes neural machine translation brings along is that the overall language quality is much better when compared to the SMT world. This does not mean that the translation is always perfect. As stated by one of our testers: if it is right, then it is astonishingly good quality. The ratio of good and poor translation naturally varies depending on the engine, but good engines can provide about 50% (or even higher) of really good translation target text.

Here are some examples showcasing what NMT post-editors can expect:

DE original:

Der Rechnungsführer sorgt für die gebotenen technischen Vorkehrungen zur wirksamen Anwendung des FWS und für dessen Überwachung.

Reference human translation:

The accounting officer shall ensure appropriate technical arrangements for aneffective functioning of the EWS and its monitoring.

Globalese NMT:

The accounting officer shall ensure the necessary technical arrangements for theeffective use of the EWS and for its monitoring.

As you can see, the output is fluent, and the differences are just preferential ones, more or less. This is highlighting another issue: automated quality metrics like BLEU score are not really sufficient to measure the quality. The example above is only a 50% match in the BLEU score, but if we look at the quality, the rating should be much higher.

Let’s look another example:

EN original

The concept of production costs must be understood as being net of any aid but inclusive of a normal level of profit.

Reference human translation:

Die Produktionskosten verstehen sich ohne Beihilfe, aber einschließlich eines normalen Gewinns.

Globalese NMT:

Der Begriff der Produktionskosten bezieht sich auf die Höhe der Beihilfe, aber einschließlich eines normalen Gewinns.

What is interesting here that the first part of the sentence sounds good, but if you look at the content, the translation is not good. This is an example of a fluent output with a bad translation. This is a typical case in the NMT world and it emphasizes the point that post-editors must examine NMT output differently than they did for SMT – in SMT, bad grammar was a clear indicator that the translation must be post-edited.

Post-editors who used to proof and correct SMT output have to change the way they are working and have to be more careful with proofreading, even if the NMT output looks alright at first glance. Also, services related to light post-editing will change – instead of correcting serious grammatical errors without checking the correctness of translation in order to create some readable content, the focus will shift to sorting out serious mistranslations. The funny thing is that one of the main problems in the SMT world was weak fluency and grammar, and now we have good fluency and grammar as an issue in the NMT world…

And finally:

DE original:

Aufgrund des rechtlichen Status der Beteiligten ist ein solcher Vorgang mit einer Beauftragung des liefernden Standorts und einer Berechnung der erbrachten Leistung verbunden.

Reference human translation:

The legal status of the companies involved in these activities means that this process is closely connected with placing orders at the location that is to supply the goods/services and calculating which goods/services they supply.

Globalese NMT:

Due to the legal status of the person, it may lead to this process at the site of the plant, and also a calculation of the completed technician.

This example shows that unfortunately, NMT can produce bad translations too. As I mentioned before, the ratio of good and bad NMT output you will face in a project always depends on the circumstances. Another weak point of NMT is that it currently cannot handle the terminology directly and it acts as a kind of “black box” with no option to directly influence the results.

Reference: https://bit.ly/2hBGsVh

Print Friendly, PDF & Email
DQF: What is it? and How it works?

DQF: What is it? and How it works?

What does DQF stand for?

DQF stands for the Dynamic Quality Framework. Quality is considered Dynamic as translation quality requirements change depending on the content type, the purpose of the content and its audience.

Why is DQF the industry benchmark?

DQF has been co-created since January 2011 by over fifty companies and organizations. Contributors include translation buyers, translation service providers, and translation technology suppliers. Practitioners continue to define requirements and best practices as they evolve through regular meetings and events.

How does DQF work?

DQF provides a commonly agreed approach to select the most appropriate translation quality evaluation model(s) and metrics depending on specific quality requirements. The underlying process, technology and resources affect the choice of quality evaluation model. DQF Content Profiling, Guidelines and Knowledge base are used when creating or refining a quality assurance program. DQF provides shared language, guidance on process and standardized metrics to help users execute quality programs more consistently and effectively. Improving efficiency within organizations and through supply chains. The result is increased customer satisfaction and a more credible quality assurance function in the translation industry.

The Content Profiling feature is used to help select the most appropriate quality evaluation model for specific requirements. This leads to the Knowledge base where you find best practices, metrics, step-by-step guides, reference templates, and use cases. The Guidelines are publicly available summaries for parts of the Knowledge base as well as related topics.

What is included in DQF?

1. Content Profiling and Knowledge base

The DQF Content Profiling Wizard is used to help select the most appropriate quality evaluation model for specific requirements. In the Knowledge Base you find supporting best practices, metrics, step-by-step guides, reference templates, use cases and more.

2. Tools

A set of tools that allows users to do different types of evaluations: adequacy, fluency, error review, productivity measurement, MT ranking and comparison. The DQF tools can be used in the cloud, offline or indirectly through the DQF API.

3. Quality Dashboard

The Quality Dashboard is available as an industry-shared platform. In the dashboard, evaluation and productivity data is visualized in a flexible reporting environment. Users can create customized reports or filter data to be reflected in the charts. Both internal and external benchmarking is supported, offering the possibility to monitor one’s own development and to compare results to industry highs, lows and averages.

4. API

The DQF API allows users to assess productivity, efficiency and quality on the fly while in the translation production mode. Developers and integrators are invited to use the API and connect with DQF from within their TMS or CAT tool environments.

References: Taus

Print Friendly, PDF & Email
Localizing Slogans: When Language Translation Gets Tricky

Localizing Slogans: When Language Translation Gets Tricky

A slogan. It seems pretty straightforward. Translating a few words, or even a sentence, shouldn’t be all that complicated, right?
And yet we’ve seen countless examples of when localizing slogans has gone awry—from big global brands—illustrating just how tricky translating slogans can be.
Anybody recall Pepsi’s “Come alive with the Pepsi generation” tagline being translated into “Pepsi brings your ancestors back from the grave” in Chinese?
While humorous, this language translation misfortune can be costly—and not just in a monetary sense. We’re talking time-to-market and brand reputation costs, too.

Why slogans pose language translation difficulties

The very nature of slogans makes them challenging to translate. Many times slogans are very creative, playing on cultural idioms and puns.
There often isn’t a direct translation that can take on the exact meaning of your slogan. And, in fact, linguists may experience translation difficulties in attempting to complete the translation word for word.
Local nuances come into play as well. Some words may have entirely different meanings than your source language and can be misinterpreted. Just think of product names that are often used in slogans. The Chevy Nova name was criticized in Latin America because “Nova” directly translates into “doesn’t go.”
Also, different cultures have unique emotional reactions to given words. Take McDonald’s and its famous slogan “I’m lovin’ it.” The fast food mogul localized this slogan to “Me encanta” or “I really like it,” so the mantra was more culturally appropriate for Spanish-speaking countries, where love is a strong word and only used in certain situations.
Because of the language translation difficulties involved, you may need a more specialized form of translation to ensure that your slogan makes a positive impact in your international markets.

How to approach localizing slogans

First and foremost, communication is vital throughout the entire localization process. When approaching slogans, we’ll collaborate with your marketing experts—whether internal or outside creative agencies—as well as your in-country linguists with marketing expertise.

Having in-country linguists’ work on your slogan is absolutely critical. These language translation experts are fully immersed in the target culture. They are cognizant of cultural nuances, slang and idioms, which ensures that your slogan will make sense—and go over well—in your target locales.

We’ll review the concepts in the tagline or slogan as a team and identify any challenging words or phrases and assess how to approach it. Oftentimes, a direct translation won’t work. We may need to localize it in a way that’s more appropriate, such as the McDonald’s “Me encanta” example above.

If it poses much difficulty, then we may need to turn to transcreation services.

Transcreation process and your slogan

The transcreation process is a specialized version of language translation that’s a highly involved and creative process.

Copywriter linguists will identify your brand qualities and portray those in a way that perfectly resonates with your target audience. Think of it as a mix of “translation” and “creation.” It’s not a word-for-word translation, but rather re-creating an idea or message so it fosters an emotional connection in a different culture.

Looking at a quick example, Nike’s celebrated slogan “Just do it” had no meaningful translation in Chinese. So instead, the message was transcreated to mean “Use sports” or “Have sport,” which had a more prominent impact in that culture.

Localizing slogans, or more specifically, your slogan, correctly can mean a stronger global brand reputation—driving revenue and increased market share worldwide. Taking a hasty, nonchalant approach can mean just the opposite. And you may find yourself having to spend time and resources rectifying what comes with a language translation error.

 Reference: https://bit.ly/2GSx36x
Print Friendly, PDF & Email
Edit Distance in Translation Industry

Edit Distance in Translation Industry

In computational linguistics, edit distance or Levenshtein distance, is a way of quantifying how dissimilar two strings (e.g., words) are to one another by counting the minimum number of operations required to transform one string into the other.  The edit distance between (a, b) is the minimum-weight series of edit operations that transforms a into b. One of the simplest sets of edit operations is that defined by Levenshtein in 1966 which are:

1- Insertion.

2- Deletion

3- Substitution.

In Levenshtein’s original definition, each of these operations has unit cost (except that substitution of a character by itself has zero cost), so the Levenshtein distance is equal to the minimum number of operations required to transform a to b.

For example, the Levenshtein distance between “kitten” and “sitting” is 3. A minimal edit script that transforms the former into the latter is:

  • kitten – sitten (substitution of “s” for “k”).
  • sitten –  sittin (substitution of “i” for “e”).
  • sittin –  sitting (insertion of “g” at the end).

What are the application of edit distance in translation industry?

1- Spell Checkers

Edit distance is applied where automatic spelling correction can determine candidate corrections for a misspelled word by selecting words from a dictionary that have a low distance to the word in question.

2- Machine Translation Evaluation and Post Editing

Edit distance can be used to compare a postedited file to the machine translated output that was the starting point for the postediting. When you calculate the edit distance, you are calculating the “effort” that the posteditor made to improve the quality of the machine translation to a certain level. Starting from the source content and same MT output, if you perform a light postediting and a full postediting, the edit distance for each task will be different, and the human quality level is expected to have a higher edit distance, because more changes are needed. This means that you are measuring light and full postediting using the edit distance.

Therefore, the edit distance is a kind of “word count” measure of the effort, similar in a way to the word count used to quantify the work of translators throughout the localization industry. It also helps in evaluating the quality of MT engine by comparing the raw MT to the post edited version by a human translator.

3- Fuzzy Match

In translation memories, edit distance is the technique of finding strings that match a pattern approximately (rather than exactly). Translation memories provide suggestions to translators, and fuzzy matches are used to measure the effort made to improve those suggestions.

Print Friendly, PDF & Email
LookAhead Feature – Towards Faster Translation Results

LookAhead Feature – Towards Faster Translation Results

 

 

To facilitate your work on SDL Trados Studio 2017 SR1, SDL powered it with LookAhead feature. LookAhead is an in-memory lookup and retrieval mechanism which ensures that your translation search results are displayed fast when you activate a segment for translation. LookAhead technology radically improves the retrieval speed of TM search results, especially for long or complex source text. Once your source text is loaded in SDL Trados Studio , the application starts matching source text strings against the available translation resources (TMs, termbases or machine translation) in the background for the next two segments after the current one. As a result, you are instantly provided with translation hits for each segment that has matching translation results.

How to enable LookAhead?

  1. Go to File, and select Options.
  2. In the Options dialog, in the navigation tree, expand Editor.
  3. Select Automation.
  4. Under Translation Memory, select the Enable LookAhead checkbox.

Print Friendly, PDF & Email
Adaptive MT – Trados 2017 New Feature

Adaptive MT – Trados 2017 New Feature


SDL Trados Studio 2017 includes new generation of machine translation.

How does it work?

It allows users to adapt SDL Language Cloud machine translation with their own preferred style. There is a free plan and it offers these features:

  • 400,000 machine translated characters per month.
  • only access to the baseline engines, so this means no industry or vertically trained engines.
  • 5 termbases, or dictionaries, which can be used to “force” the engine to use the translation you want for certain words/phrases.
  • 1 Adaptive engine.
  • Translator… this is basically a similar feature to FreeTranslation.com except it’s personalized with your Engine(s) and your termbases.

How does it help?

  • Faster translation with smarter MT suggestions.
  • Easy to use and get started.
  • Completely secure – no data is collected or shared.
  • Unique MT output, personal to you.
  • Access directly within Studio 2017.
  • No translation memory needed to train the MT.
  • Automatic, real time learning – no pre-training required.

What are the available language pairs?

Uptill now, Adaptive MT is available in these language pairs:

English <-> French
English <-> German
English <-> Italian
English <-> Spanish
English <-> Dutch
English <-> Portuguese
English <-> Japanese
English <-> Chinese

For reference: https://www.sdltrados.com/products/trados-studio/adaptivemt/

Print Friendly, PDF & Email