Machine Translation
Cost-effective.
Expedited timelines.
Technology powered.
Human editors.
Quality translations.
Machine Translation with Post-Editing (MTPE) is a hybrid approach to translation that harnesses the speed of machine translation while ensuring the accuracy, style, and nuances that only a professional, human translator can deliver.
Why Utilize Machine Translation with Post-Editing?
A number of factors, including new language access regulations, are increasing organizations’ need for professional translations. Unfortunately, budgets rarely keep up accordingly to support these needs. Many are turning to machine translation for an affordable solution. Alone, however, this automation can introduce risk.
Machine translation engines are built, expanded, and refined based on statistical or rule-based algorithms to translate from one language to another. That’s a lot of technical speak for tools that can provide a sufficient baseline for straightforward text, but often fall short when it comes to specialized terminology, acronyms, idioms, and nuances in dialect.
By coupling machine translation tools with a review by a professional translator, Alboum’s MTPE provides organizations with several benefits:
- Speed: Quicker turnaround times compared to traditional human-only translation.
- Cost-Effectiveness: Lower costs while maintaining quality standards.
- Consistency: Uniformity in terminology and style across large volumes of content.
- Scalability: Ability to handle large-scale translation projects efficiently.
- Customization: Nuanced language appropriate for your industry and specific audience, as well as that considers your organization’s preferred terms.
Which Machine Translation Tool is Used?
Many organizations make use of automated translation tools built into existing systems and platforms. Others utilize tools like ChatGPT or Google Translate to do an initial machine translation. The quality of these free tools varies widely.
Alboum’s MTPE draws on multiple translation engines that work within Trados, our FERPA-compliant translation memory tool. We also leverage all other translation work we’ve completed for your organization, as well as your translation glossaries and style guides, to support the editing phase of our multi-step approach.
Security
Publicly available translation tools like Google Translate and the free version of ChatGPT lack any form of security. When you put information into Google Translate, be it a student’s IEP plan or patient’s medical history, that information becomes part of Google’s database.
In contrast, as Alboum’s machine translation engine is built into our already-HIPAA- and FERPA-compliant environment, your content is secure, archived, and deleted according to existing policies.
Language Trends
Languages evolve every day, incorporating new slang, colloquial speech, and new terms, as well as adapting to incorporate social and cultural advancements like gender-neutral terms.
While AI-supported translation tools are always learning, they don’t consider your specific community, brand guidelines, or terminology preferences. Alboum draws upon previously approved translations, as well as integrates your preferences over the automatic word choice.
The result is a significantly higher quality translation that’s in line with your existing tone and style.
Why Do I Need an Editor at all?
Machine translation tools have made significant strides in recent years, and are a good option when you need a quick option to understand what something means. They are an appropriate choice for a word-for-word translation of an isolated word or two when context or cultural nuance isn’t critical.
Yet, we’ve all seen examples of translations that were clearly not done by a native speaker. While these translations may be understandable, they don’t inspire confidence nor do they help engage an audience.
A professional linguist reviews the machine output for completion, consistency, and cohesion. They correct mistranslations, literal translations, grammatical errors, and misspellings, as well as ensure all content is translated and culturally appropriate. The result is a natural-sounding overall tone of the text and accurate meaning of each phrase rather than a word-for-word translation.
Is MTPE Available in All Languages?
Languages that have more publicly-available content have better engines because they simply have more to draw on. Spanish, French, and German are examples of commonly translated languages in which machine translations perform well.
By contrast, the engines for languages like Dari, Hebrew, and Somali just aren’t evolved enough to generate a good product so post-editing can sometimes take longer and end up being more costly than translating from scratch.
When Should I Opt for Machine Translation with Post-Editing?
Free Built-In and Online Translation Tools
Alboum MT-PE
Situations it’s best for
- Getting a general sense of what something means or someone is saying.
- Short, simple sentences or isolated words without context.
- Long documents that need to be translated on a tighter timeline.
- Documents with straightforward language.
- Documents that need an accurate translation completed on a budget.
Situations it’s NOT recommended for or the approach introduces risk
- Documents that legally require translation or set forth policies or protocols.
- Medical, technical or industry-specific content, particularly patient instructions.
- Gender-neutral language and newer terms that may be adopted in one language but not another.
- Content that includes cultural nuance, subtle intonations, or conversation, such as books.
- Social media and marketing/creative content that is typically highly nuanced.
- Super short text.
Security Concerns
Free online tools use submitted content to train engines. Built-in tools require confirmation of security protections and confirmation whether translations are completed within the tool or sent back and forth to outside engines.
Content is processed only within Alboum’s HIPAA- and FERPA-compliant platforms and deleted/archived according to our security schedules or sooner on request.
Style and tone
Word-for-word translation that draws on whatever materials are published online using the same words.
Prioritizes your existing translation glossaries and draws on translation memory from documents previously translated by Alboum. Editing by human beings who understand how words are used in context.
Final product
What you see is what you get.
A professional linguist reviews the machine translation output to correct mistakes, mistranslations, and missing segments of text common with this type of technology.
Linguists also review the original text to confirm the meaning is accurately reflected and any necessary corrections are made to ensure cultural relevance.
Quality of machine translation engine
Anyone online can suggest translations that the system will automatically incorporate, even if they are incorrect. As tools increasingly rely on AI, hallucinations and inaccuracies present real risk.
Accuracy and quality of machine translation output is reviewed regularly based by a team of professional linguists. Your organization’s specific termbase is updated with each translation Alboum completes.
